SINGAPORE (ANN/THE STRAITS TIMES) – Singapore’s National Supercomputing Centre (NSCC) will receive a SGD270 million investment over the next three years from the National Research Foundation (NRF) to expand its capabilities.
The funding will focus on developing next-generation supercomputers tailored for artificial intelligence (AI) applications and on fostering a skilled workforce in supercomputing.
This grant represents the second tranche of funding from the NRF for supercomputers, which can perform in hours complex calculations that would take standard computers days to complete.
In 2019, the NRF invested SGD200 million to support the projects and initiatives by NSCC Singapore, which was established in 2015 to manage the Republic’s high-performance computing and drive research.
In Singapore, there is a shortage of specialists with the skills needed to operate supercomputers, which use more complex algorithms to compute data at a larger scale, NSCC chief executive Terence Hung told the media on October 25 at the launch event of the centre’s new supercomputers.
It can take upwards of a year for programmers to come to grips with supercomputers, said Dr Hung, who estimated that there were fewer than 100 professionals proficient in the technology here.
Through collaborations with academia, workshops and initiatives that will tap the SGD270 million grant, he aims to grow the numbers of tech talent proficient in supercomputing into the thousands over the next few years.
“We have users who can run applications, but we need users who can run it effectively,” he said. “That needs training, and we are definitely short on that right now.”
Deputy Prime Minster and NRF chairman Heng Swee Keat, while announcing the grant at the NSCC premises located within NUS on October 25, said developing talent was crucial to reap the benefits of supercomputing, which can lead to breakthroughs in research in AI, healthcare and climate change, among other endeavours.
Supercomputers have found solutions to many challenges, such as enhancing the accuracy of weather predictions, supporting drug discovery during the Covid-19 pandemic and improving urban planning through traffic simulations, he said.
DPM Heng unveiled NSCC’s latest supercomputers, the Aspire 2A+ and 2A – huge room-sized machines with up to 200,000 times the computing power of an average laptop, operating from a data centre located at the NSCC and which clients can remotely access.
The SGD40 million Aspire 2A serves as the go-to supercomputer for the majority of researchers, Dr Hung said.
Its more advanced brother, the 2A+, contains 320 sets of the latest Nvidia H100 graphics processing units and specialises in AI research and large language models in the same vein as ChatGPT, which runs on heaps of data.
DPM Heng said: “(The new supercomputers) will significantly advance computing capabilities that better support our policymakers and researchers in addressing emerging problems and working with global partners to tackle shared challenges.”
The supercomputers have been used to analyse 3,000 years of climate change data to predict changes to Singapore’s sea levels – information that the authorities can use to guide urban planning, said Professor Dale Barker, director of the National Environment Agency’s Centre for Climate Research.
The research, supported by supercomputers, has generated detailed forecasts predicting more extreme rainfall, hotter temperatures and rising sea levels.
The latest NSCC computers expand the pool of supercomputers available to Singapore researchers, following collaborations with Japan, Taiwan and Australia to share their supercomputers.
The technology, which is compatible with standard computers, is still relied on by researchers to develop algorithms, even as developers worldwide race towards quantum computers, the next generation of supercomputers.
Advances in supercomputing will also support Singapore’s SGD120 million AI for Science initiative, announced by Prime Minister Lawrence Wong during the launch of Singapore’s renewed Smart Nation 2.0 vision in October, to develop AI capabilities in scientific research.
