'Fastest' AI supercomputer in academia to work on climate change, coronavirus projects
The University of Florida (UF) has unveiled a $70 million partnership with Nvidia to boost the institution's AI capabilities thanks to what it claims will be the fastest supercomputer in academia.
The deal will see UF upgrade its existing supercomputer, the HiPerGator 3.0, from maximum speeds of just over one petaflop to 700 petaflops of AI performance, which the institution hopes will advance academic research at an unprecedented pace.
Artificial Intelligence
To boost the capabilities of the HiPerGator, UF will deploy Nvidia's DGX A100 systems, which are designed to build and run AI projects at a large scale. It will be the first time an institution of higher learning in the US accesses the A100 systems.
SEE: Managing AI and ML in the enterprise 2020: Tech leaders increase project development and implementation (TechRepublic Premium)
The existing supercomputer will integrate 140 A100 systems powered by 1,120 A100 Tensor Core GPUs, and will include four petabytes of high-performance storage. The overall infrastructure comes as a single software-defined platform that covers data analytics, AI training and inference acceleration and recommendation systems.
Nvidia CEO Jensen Huang said: "We are excited to contribute Nvidia's expertise to work together to make UF a national leader in AI and help address not only the region's, but the nation's challenges."
The deal comes as researchers increasingly tap into supercomputing power to address the issues caused by the COVID-19 crisis. The US government, in partnership with IBM, recently announced the launch of a COVID-19 High Performance Computing Consortium, granting researchers across the globe access to supercomputers in the US to model the disease and improve understanding of the pandemic.
UF anticipates that, when set up, the new capabilities will contribute to researching solutions to pandemic-related challenges, but not only that. Supercomputing tools are also expected to enable academia to apply AI to address issues such as rising seas in a state that has more than 1,300 miles of coastline; as well as aging populations, data security, personalized medicine, urban transportation, and food insecurity in a time of climate change.
SEE: ARM and Linux take the supercomputer TOP500 crown
Speaking about UF's partnership with Nvidia, Florida's governor Ron DeSantis said: "As we look to keep our best talent in state, this will be a significant carrot, you'll also see people around the country want to come to Florida."
UF joins other US universities in deploying top-level supercomputing power to support AI research. Last year, the Rensselaer Polytechnic Institute in New York revealed that it was using AiMOS, IBM's latest eight-petaflop supercomputer, to advance the development of stronger hardware that can support new AI applications.
Back in Florida, the HiPerGator is expected to complete its upgrade by early 2021, a few weeks after Nvidia's A100 system is delivered.