<\/p>\n
A supercomputer at the National University of Singapore (NUS) has become the first from the city-state to make it into the coveted Top500 list<\/a> of the world\u2019s most powerful supercomputers.<\/p>\n Dubbed Hopper, the supercomputer was placed 105th on the most recent list, which was released in June 2025. It is also Singapore\u2019s highest-ranked supercomputer<\/a> and the only one managed by a university in Southeast Asia to make the cut, a significant achievement for the country\u2019s academic research capacity.<\/p>\n Hopper is capable of performing 25 quadrillion calculations per second, or 25 petaflops. For comparison, a researcher performing one calculation per second would need millions of years to accomplish what Hopper performs in just one second.<\/p>\n That amount of compute capacity is already being used to support ambitious, data-intensive research projects across the university that were previously constrained by time or computing power.<\/p>\n One such project focuses on next-generation battery design for clean energy<\/a>. Deng Zeyu, an assistant professor from NUS\u2019s department of materials science and engineering, is using Hopper to do large-scale quantum mechanical calculations and complete complex simulations up to 12 times faster than with traditional systems.<\/p>\n \u201cTasks that used to take three months now take a week,\u201d said Deng. \u201cIt\u2019s like switching from a bicycle to a high-speed train \u2013 Hopper has completely transformed our research pace and possibilities, allowing us to explore new frontiers in battery technology.\u201d<\/p>\n In the realm of artificial intelligence<\/a> (AI), assistant professor Mike Shou from NUS department of electrical and computer engineering described Hopper as a \u201cgame-changer\u201d for his research into developing new methods for machines to understand actions and complex events in videos.<\/p>\n \u201cIts powerful performance and system stability have laid a solid foundation for our work in video generative AI<\/a>,\u201d he said. \u201cWe can now generate both short-form and long-form videos and build multimodal agents that operate across a wide range of previously unseen environments.\u201d<\/p>\n Shou\u2019s work has the potential to fuel applications ranging from perception systems for self-driving cars to intelligent video creation tools for journalists and filmmakers.<\/p>\n Hopper is also driving advances in biomedical engineering. SurgVLM, a vision-language model meant to facilitate surgical intelligence, was developed by assistant professor Jin Yueming and her colleagues from the university\u2019s department of biomedical engineering using the supercomputer.<\/p>\n \u201cHopper\u2019s Nvidia H100<\/a> GPU [graphics processing unit] nodes and the high-speed InfiniBand network<\/a> have been instrumental to our key experiments,\u201d said Jin, noting that the model could one day provide real-time cognitive assistance during surgery and aid in robotic procedures.<\/p>\n