March 10, 2025, The preliminary results of the 2025 ASC Student Supercomputer Challenge (ASC25) were officially announced. All three teams from our university won the Second Prize, achieving a significant rise in overall rankings compared to previous years. This marks our best-ever performance in the ASC competition, and showcases the students’ strong learning and practical abilities in supercomputing application analysis, artificial intelligence algorithms, performance optimization, and parallel computing strategy design.

These supercomputing teams of our univerisity were spearheaded by the Department of Artificial Intelligence, School of Computer Science, with instructors Peng Cheng, Wan Tang,  Xiao Zhang, and Ximin Yang providing guidance. The teams were composed of undergraduates from various majors. Under the comprehensive guidance of instructor group, the participating teams carried out a year-long integrated training program centered on "Theory, Practice, and Innovation," leveraging resources from the Wuhan Artificial Intelligence Computing Center, Modern Educational Technology Center, and the Central Laboratory of the School of Computer Science.

The training introduced an mentorship model, where experienced former core members — Jiayi Lan (Major in Artificial Intelligence), Suting Li (Major in Network Engineering), and Kewei Chen (Major in Software Engineering) — formed a technical think tank. They constructed a specialized knowledge graph tailored to ASC competition features to provide systematic technical support for new team members. The three teams included members from Artificial Intelligence, Computer Science and Technology, Software Engineering, and Information Science majors, enabling effective interdisciplinary collaboration in algorithm design, system tuning, and mathematical modeling, thereby fully demonstrating the innovative potential of cross-disciplinary integration.

This year’s competition featured, for the first time, questions inspired by the latest Nobel Prize-winning research, focusing on the convergence of high-performance computing and artificial intelligence in frontier scientific fields. The RNA methylation site detection task required optimizing the performance of a detection pipeline for "5-methylcytosine" (m5C) modifications on mRNA. Teams had to sort out the process steps, build the software environment, and process approximately 340 million sequencing data entries across three datasets, optimizing performance while ensuring accuracy. Another challenge involved inference optimization for AlphaFold3. Teams were tasked with optimizing the structure prediction code of AlphaFold3 and migrating it from GPU to CPU platforms. They needed to complete efficient predictions on twelve designated protein sequences while maintaining accuracy.

The ASC Student Supercomputer Challenge is the world’s largest supercomputing competition for university students. It aims to foster global collaboration and talent development among young HPC professionals, enhance the application and research capabilities, and promote technological innovation and industrial advancement through the power of supercomputing. This year, over 300 teams participated in the preliminary round, and 25 finalist teams including those from Zhejiang University, Peking University, Tsinghua University, University of the Witwatersrand, and Friedrich-Alexander-University Erlangen-Nuremberg will advance to the on-site finals in May.