When a supercomputer encounters quantum computing, what kind of sparks will it produce?

Not long ago, the supercomputing application team from China won the “Gordon Bell” award, the highest international award in the field of supercomputing applications, by virtue of the quantum computing simulator-"super-large-scale quantum random circuit real-time simulation" project.

　　What is less known is that behind the Supercomputing Awards is a multi-disciplinary, youthful and energetic Chinese "dream team."

The team members include 9 members including Liu Yong and Liu Xin from Zhijiang Laboratory and the National Supercomputing Wuxi Center, Fu Haohuan and Chen Dexun from Tsinghua University and the National Supercomputing Wuxi Center, and Wu Wenzhao from the National Supercomputing Wuxi Center. 14 members including Huang Heliang and Guo Chu of the Shanghai Quantum Science Research Center had an average age of 34 years old at the time of the award.

Among them, Liu Xin, Fu Haohuan, Guo Chu, and Chen Dexun are the co-corresponding authors of the result paper.

　　Fu Haohuan, the co-corresponding author of the paper, professor of the Department of Earth System Sciences of Tsinghua University, and deputy director of the National Supercomputing Wuxi Center, told a reporter from China Youth Daily and China Youth Daily that the award was given to the interdisciplinary team’s “strong alliance” "The power of ", the majors involved span multiple disciplines such as computer, mathematics, physics, and quantum computing.

　　"What we have done this time is an application of supercomputers for quantum computing. This team includes people who study supercomputing, people who study quantum, and experts in mathematics and physics. The level of expertise to do integrated innovation. I think this is also an important reason for the final award." Fu Haohuan said.

　　Quantum computing has its own unique performance advantages. Scientists in many countries are trying to develop quantum computers.

However, with the expansion of the scale of quantum computers, how to verify the accuracy of quantum computers has become one of the main bottlenecks in the development of quantum computers.

This latest research by the China Supercomputing Application Team, based on previous work, proposes a set of its own unique algorithm to achieve quasi-real-time quantum simulation, bringing people one step closer to the birth of quantum computers in the future.

　　In some application scenarios, quantum computing devices can show performance advantages that traditional computers cannot match.

This is also the "quantum hegemony" often referred to in the scientific community.

　　"Why do we have to use a classical computer to verify? This is because to make a quantum computer, we must first ensure the accuracy of the calculation results. However, the running time of a quantum computer is measured in seconds, and verification by a classical computer takes days. ', even in units of'year', so one of the main difficulties in quantum computer research and development now lies in how to find a good algorithm to make the verification work faster." One of the team members, from the Shanghai Quantum Science Research Center Guo Chu explained.

　　In order to overcome such large and complex obstacles, the team has worked on multiple levels including quantum simulation algorithms, tensor network graph segmentation methods, shrinking path optimization, and tensor operator optimization related to the underlying supercomputer hardware. The application is truly applied. The integration of algorithms and architecture has realized the quasi-real-time quantum circuit simulation based on the new generation of Shenwei Supercomputing.

　　"I think multidisciplinary cross-discipline plays a very important role in our work, and it is an important reason why research can progress smoothly." As a young researcher who studies quantum, Guo Chu, 32, made a "crossover" during his Ph.D. Let the two unrelated fields of supercomputing and quantum computing collide.

　　"During my Ph.D., as a physics student, I may write code far more than a computer science student." Guo Chu said that during his Ph.D., he spent a year and a half programming himself and implemented it with code. A set of tensor network algorithms.

He said, “In fact, I was doing a relatively basic job at that time, which took a long time, and therefore I couldn’t publish as many papers as my classmates. Even this set of codes was during Ph.D. and even after graduation. I haven't used it either."

　　However, opportunities are always reserved for those who are prepared. In this scientific research result, Guo Chu's once seemingly "useless" has become one of the core technologies.

　　Speaking of his experience in the field of supercomputing and even winning international awards, Guo Chu said with a smile: "It was a coincidence to be able to work with the supercomputing team. We happened to meet at an academic conference in 2019. , Chatting with each other about their respective research and understanding their respective strengths and needs, only to find opportunities for cooperation. In fact, we are two completely unrelated fields."

　　In the team, Guo Chu is already considered an "old man". He said that the young people in the team will always be the "Desperate Saburo" fighting at the forefront, full of enthusiasm and drive.

　　As one of the team members, Wu Wenzhao used a whole night to implement a good algorithm on the machine, which directly increased the computing efficiency of the entire project by more than 10 times, which impressed Guo Chu.

　　Wu Wenzhao was a master student in the Department of Earth Science of Tsinghua University. After graduation, he worked at the National Supercomputing Wuxi Center. Since 2020, he has been responsible for the development of a new generation of Shenwei supercomputing platform deep learning framework.

"During my studies in the Department of Earth Sciences, the department has always encouraged cross-discipline. Now, I am honored that I can continuously try to apply the knowledge I have learned to new fields." said Wu Wenzhao.

　　This award is also the third time that the Tsinghua University supercomputing team has won the "Gordon Bell" award.

From 2016 to 2017, the atmospheric dynamic equation solver and the non-linear seismic simulation work led by the Tsinghua Supercomputing team respectively achieved "zero" breakthroughs and re-elections in this award in the past 30 years.

　　As a core member of the team, Fu Haohuan said that "crossing boundaries" has become a daily routine for his work.

　　"We can't supercompute for the sake of supercomputing. We have a good hardware foundation. Ultimately, we have to make supercomputing truly serve all aspects of economic and social development through interdisciplinary methods, and this can truly reflect the supercomputing Value." Fu Haohuan said.

　　China Youth Daily·China Youth Daily reporter Ye Yuting Source: China Youth Daily