HKBU team wins First Class Award in international student supercomputer challenge
10 Jul 2018
A team of five Computer Science students of HKBU won the First Class Award in the ASC Student Supercomputer Challenge 2018 organised by the Asia Supercomputer Community held recently. Among the 300 teams from around the world, the HKBU team secured a place in the final round of the Challenge and won the First Class Award. The HKBU team was the only Hong Kong team to enter the final round.
Led by Dr Chu Xiaowen, Associate Professor of the Department of Computer Science, the team comprised Year 2 student Ni Ronghao and Year 3 students Wang Shihao, Feng Zijin, Wang Haixin and Wong Tsz-shing.
Team member Wong Shihao said, “Our team members learned state-of-the-art technologies used in supercomputers and artificial intelligence through the competition. In the process, we encountered huge difficulties in both hardware and software aspects. For instance, during the final round of the Challenge, our programming communication encountered a fatal error, a problem that was beyond our expectation despite our practice prior to the competition. We managed to replace the affected application with another interface, making the cluster work again.” Shihao expressed his gratitude to the team supervisor Dr Chu Xiaowen and assistant coach to the team Mr Liu Chengjian who offered the team a lot of advice and assistance to successfully conquer each problem step by step.
The Challenge is one of the world’s largest supercomputing contests. After two months of preliminary matches, 20 teams were selected for the final round of the Challenge, in which teams had to design and build their own supercomputing system, with the limit of 3,000 watts of power, and apply it to six scientific and engineering projects. They were also requested to develop a system to optimise the 3-dimensional reconstruction algorithm used in Cryo-Electron Microscopy, an imaging method which won the 2017 Nobel Prize in Chemistry. For another artificial intelligence challenge, each competing team applied deep learning and high-performance computing techniques to enable the system to comprehend text and answer users’ questions.