Despite the ongoing COVID-19 pandemic, more than 300 teams from around the world participated in the preliminaries of ASC20-21. The competition demonstrated the contestants’ profound understanding of supercomputing technologies and application optimization.
The preliminary results for the 2020-2021 ASC Student Supercomputer Challenge (ASC20-21) have now been released. After two months of intense competition, 28 teams from over 300 participating universities from around the world have stood out and advanced to the finals. Beihang University, Jinan University, Shanghai Jiaotong University, the Chinese University of Hong Kong, Universidad EAFIT, University of Warsaw, and Ural Federal University are just some of the finalists.
Because the COVID-19 pandemic is still rampant in many countries around the world, the ASC20-21 Student Supercomputer Challenge finals will be organized to include both on-site and virtual participations. The Top21 teams from the Chinese mainland will join the finals on-site, and the Top7 teams outside of Chinese mainland will remotely participate in the finals virtually.
Among the teams advancing to the ASC20-21 finals are three champions from the prior Student Cluster Competitions: Tsinghua University - champion of SC19 & SC20, University of Science and Technology of China - champion of ISC20, and National Tsing Hua University - champion of ASC19. It will be the first time in the history of the student supercomputing competition that three world champions will directly compete, which has created massive interest and enthusiasm for this round of the ASC20-21 finals.
There is also much excitement about the debuting teams from Monash University, Hunan University and Lanzhou University for making their first ASC finals appearance this year. Five teams, including the team of Peking University, are repeat finalists after advancing to their first finals at ASC19. This ASC20-21 final is full of traditional powerhouses and remarkable new contenders, which will make the competition intense till the last moments, when various awards will be presented including the Champion, Silver Prize, Highest LINPACK, and e Prize.
The ASC20-21 preliminaries included computational tasks from ACS20 like quantum computing simulation and language examination, but also included a new cutting-edge task in astronomy -- searching for pulsars. Many of the participating teams performed very well in the preliminaries and demonstrated an outstanding ability to learn and innovate freely.
The pulsar search task required teams to use the open-source software PRESTO to search for potential pulsars based on the observational data from the Five-hundred-meter Aperture Spherical radio Telescope (FAST), also known as "China Sky Eye". The research of pulsars is so valuable because it can help solve major physics conundrums, such as gravitational wave detection and spacecraft navigation. The Universidad EAFIT team optimized the software compilation parameters, used a variety of Python tools for algorithm hotspot analysis, and adopted a multi-process solution to parallelize the entire search process.
The Language Exam task required all participating teams to train AI models on an English Cloze Test dataset, striving to achieve the highest "test scores". The dataset covers multiple levels of English language tests used in China. The team of the Chinese University of Hong Kong carefully analyzed the intention and objectives of this task. Using a large amount of research literature, they ran experiments to compare the performance of different models, incorporated pre & post processing algorithms, and selected the best model. This boosted them to reach their remarkably high score.
The quantum computing simulation task required each participating team to use the QuEST (Quantum Exact Simulation Toolkit) running on a computer cluster to simulate 30 qubits in two cases: quantum random circuits (random.c), and quantum fast Fourier transform circuits (GHZ_QFT.c). Quantum simulations provides a reliable platform for studying quantum algorithms, which are particularly important because quantum computers are not practically available yet in the industry. National Tsing Hua University made considerable effort in finding the algorithm’s hotspots and provided a detailed performance analysis of QuEST. They tried several methods for the GPU version of QuEST and made crucial optimizations to achieve much better performance of the toolkit. In addition, they delivered a detailed plan to further optimize QuEST in the future.
The selection process for both on-site and virtual participation has presented a tough challenge for the Evaluation Committee. After a long and careful evaluation, the ASC20-21 Student Supercomputer Challenge Committee selected the following teams to qualify for the ASC20-21 Finals:
Finalist On-site Participation
Taiyuan University of Technology
Sun Yat-sen University
National University of Defense Technology
Southern University of Science and Technology
University of Electronic Science and Technology of China
Huazhong University of Science & Technology
Shanghai Jiaotong University
University of Science and Technology of China
Northwestern Polytechnical University
Finalist Virtual Participation
National Tsing Hua University
The Chinese University of Hong Kong
University of Warsaw
Ural Federal University named after the first President of Russia B.N.Yeltsin
Note: The finalists are listed by competition ranking.
The ASC Student Supercomputer Challenge is the world’s largest student supercomputer competition, sponsored and organized by Asia Supercomputer Community and supported by Asian, European, and American experts and institutions. The main objectives of ASC are to encourage exchange and training of young supercomputing talent from different countries, improve supercomputing applications and R&D capacity, boost the development of supercomputing, and promote technical and industrial innovation. The first ASC Student Supercomputer Challenge was held in 2012 and since has attracted nearly 10,000 undergraduates from all over the world. Learn more ASC at https://www.asc-events.org/.
View source version on businesswire.com: https://www.businesswire.com/news/home/20210127005355/en/