Director General Chau-lyan Chang: Enhancing Big Data, Boosting Tech Innovation and Industrial Transformation through HPC

Photo: Director General Chau-lyan Chang emphasized the importance of data, as data will be a major key to future technology development and innovation.
 

2022/05/10 - DIGITIMES interviewed by Judy Lin (Source: This article has been extracted from DIGITIMES.)
 

High-performance computing technology has been receiving great attention in developed countries and is widely applied in technology innovation, commercial big data computing, and aerospace. National Center for High-performance Computing in Taiwan also serves academia and industry with critical high-performance computing resources via its high-speed network infrastructure and computing capabilities. The new Director General of the National Center for High-performance Computing, Chau-lyan Chang, was a senior researcher in the Computational Aerosciences Branch at NASA Langley Research Computing Center, who came back to Taiwan in April. DIGITIMES made an exclusive interview with Dr. Chau-lyan Chang to learn how he will use his NASA experience to lead NCHC and how he considers the future of high-performance computing.
 

Q: Mr. Director General, we know that you originally studied mechanical engineering, while later you changed to learn aerospace fluid dynamics and engineering software design. What made you choose the cross-disciplinary development? Why did you decide to serve in Taiwan after all you had gained in the United States?
 

I grew up in Taiwan and want to serve the country since I was young. However, I've been working for the Americans all my life, so I did not hesitate to seize the opportunity. If I come back in a few years after my retirement, I'll be too old to do anything. It's now or never.　
 

When I was a junior in the Department of Mechanical Engineering at National Taiwan University, I took a course in fluid dynamics and was recommended to read a popular science work on fluid dynamics written by A. H. Shapiro, a professor at MIT. I was fascinated by this book, in which I found that fluid dynamics was actually a perfect combination of mathematics and physics, and it solved many problems completely with mathematical models. It made me feel like a gift from above to mankind. That is why I love fluid dynamics.
 

With regard to the reason I entered the software field, I grew up in a time when computers were not well developed and I had not taken any software design courses. I had a keen interest in Fortran programming language when I studied in the research institute, with no chance to learn computer knowledge and few courses for any software design before. I borrowed a book and tried to design a program to guess four numbers. It took me a whole month to test such a program, but I was very excited when it got all the numbers correct. When I was in a research institute, I started to learn the programming language for computing. I learned it hard with a great sense of accomplishment, and that was where my interest lay. The combination of the two studies became the so-called computational fluid dynamics, which is what I mainly did when I went to the US for my Ph.D and later got a job at NASA.
 

Q: The vision of the National Center for High-performance Computing is to "become an international high-performance computing center for scientific discovery and technological innovation"; it has focused on cloud-based applications such as environment and disaster prevention, biomedicine, science and engineering, digital culture and creative industry, and research on forward-looking innovative application technologies and services such as deep learning and artificial intelligence. What do you think if there are some new areas to explore for the center to act as an enabler? What will you do to advance at the National Center for High-performance Computing? What are your further expectation and vision?
 

I want to thank the leadership of its former director general. The National Center for High-performance Computing has made significant accomplishments in IaaS, Paas, and support for Executive Yuan's technology policy, as well as in digital transformation, AI applications, and information security in the past few years. Being a leading lab in Taiwan, we want to play the role of HPC and AI technology enabler, which provides IaaS to PaaS, and also continues to strengthen network measures and introduce tools for big data analysis.
 

As for the hardware, we started building Taiwania supercomputers 1-3 in 2017, with CPU and GPU in the cluster. I would like to emphasize that the speed of these three computers has been ranked among the top 500 in the world for a period, which also demonstrates Taiwan's technological strength. In the past, the NCHC has provided platform services in environmental disaster prevention, biomedical science, and scientific engineering in cooperation with the government policies, offering invaluable high-performance computing research resources to the academic and research communities, as well as digital culture and creative activities, which was very special for high-performance computing.
 

Taiwan has several large high-performance networks, which are planned to be connected by the National Center for High-performance Computing, allowing for smoother connections between networks by solving bottlenecks. Executive Yuan's forward-looking plan includes the establishment of a 5G network, as well as a submarine cable and 5G networking center to connect the fiber-optic high-speed network from the south to the north, which can provide an alternative data transmission option to connect further south to Southeast Asia, South Asia, aiming at attracting multinational companies and international organizations to set up data centers in Taiwan. All these are where we can act as an enabler.
 

The Ministry of Science and Technology has set a goal in six major industries, including digital transformation, precision medicine, 5G/6G network, information security, and network infrastructure. In addition, we have also studied crypto technology and the method of applying IoT and blockchain to individual industries in Taiwan. We traveled around Taiwan to promote and assist vendors to adopt these technologies. Cloud services provided by international enterprises including Microsoft and Google are thriving. The National Center for High-performance Computing also leased 50% capacity of Taiwania 2 to private companies, regarded as a new successful example of collaboration with private companies, which is rarely found around the world.
 

Developing Data Strength
 

I would like to focus on the importance of data in the future, which will be a major pivot for our technology development and innovation. It can be studied from several aspects: first, for the way to collect data, for example, how does Civil IoT Taiwan use sensors to collect data and how to use biomedical data; second, for the way to produce data. Actually the application data are mainly generated by simulation. Therefore, HPC is inseparable from modeling and big data analysis, which will be one of the key issues we would explore in the future.
 

After obtaining data, what can you do to ensure data security? Therefore, we would make efforts in enhancing information security, as well as large-scale data servers and high-performance networks, to ensure a complete service system for the data chain.
 

Third, the artificial intelligence models have matured so far, with little change for many algorithms. Therefore, the quality of the data would become the key to final victory. Large amounts of consistent and perfect data are more likely to train the best models. We want to work this way to acquire data, process and protect them, and scale up the trained models in the most efficient way. The era of data will lead to the trend of the next century, and all parties would fight for data.
 

As for R&D, we would continue to act as an enabler, and further as an enhancer. We introduced many AI knowledge and tools to different industries, which proved to be helpful, and how to enhance the existing models is to be done in the future. For some industries where no data are thought to be produced, we would help them improve their productivity through the application of AI models.
 

For some of our previous success cases, the industry made a demand first and worked on a consignment basis. We would like to extend our cooperation with the industry to create more stakeholders. For example, we discovered that there were few areas to introduce high-performance computing in engineering and manufacturing, or just limited to commercial software. We hope to help them in further applications through high-performance computing simulation.
 

We also work with the academia community and start-ups. We can provide technology and integration mechanisms in IaaS, PaaS, and SaaS for big data, AI, and high-performance computing. They also have more specific expertise to complement each other to solve problems for the industry.
 

Many vendors do not know how to use HPC to design products and solve problems, nor have they thought that they need to ask the right questions for using data. Taiwan needs a paradigm shift from parts to integration into a complete solution, where the National Center for High-performance Computing can play an important role.
 

In addition, Executive Yuan just announced the establishment of Quantum Taiwan, involving the Ministry of Science and Technology, Academia Sinica to various university research units in the development of hardware, innovation of low-temperature environment technology, and software. While even with the hardware, and the function of quantum bits, another supercomputer is still needed to interpret the data, and that is the role the NCHC would act. We are also using the concept of quantum computer computing to develop encryption technology, with the aim of making a significant contribution to quantum computers in the future.
 

Q: Recently, low-earth-orbit satellite has become the focus of the information and communication industry, which are seen as a new growth opportunity for the future. You have also worked at NASA for years, can you share how NASA uses supercomputers and high-performance computing in advancing basic research and applications in the space field? How will the National Center for High-performance Computing help industries in leveraging their resources for technological innovation?
 

The United States has applied supercomputers, high-performance computing plus AI, and big data analysis in all fields, not merely space areas. We saw the news report on TV about NASA's off-road explorer traveling on the surface of Mars, so how was the Mars expedition accomplished? First, supercomputers are needed to improve the rockets' performance, reduce vibration, and the impact of carrier load during the design phase; the whole process of launching the rocket, long hours of flying, and final landing after breaking through the atmosphere requires supercomputer computing.
 

Computation and analysis are required before designing details, and detailed issues can be solved with the simulator of more sophisticated analysis. After the rocket lifts off, there is a thruster inside to control the direction, which is also calculated and simulated to ensure the rocket's heading and overall safety. After entering the atmosphere, there is also a big problem that the parachute needs to be opened when the rocket lands on Mars to ensure a safe landing, which is another different concept from the Earth. The composition of the atmosphere of Mars, mainly CO2, is different from that on Earth, but we have no way to go there first to do experiments; instead, we use computers to predict and design. Therefore, we can say that the development of the whole space industry is closely connected to the HPC simulation, and what makes the US technology lead the world? Its advantage lies in its ability to apply supercomputer computing technology to innovation, and by that United States ranks first in the world.
 

Industries in Taiwan should change the way of thinking. In America, there's no need to tell people what the high-performance computing simulation can do, which has already been a consensus there. Some people may wonder why we invest so much money in innovation. Most vendors are limited to making components, without the ambition to make machines or complete solutions. To achieve this goal, it requires system integration and simulation, which is also a necessary process for digital transformation.
 

In addition, we have some simple ideas on how to use the digital twin to solve complex problems, and we hope to contribute to it through cross-departmental collaboration.

URL of the original article from DIGITIMES:  https://www.digitimes.com/news/a20220428VL205/hpc.html
 

Brief Introduction of Director General Chau-lyan Chang: A bachelor's and master's degree in Mechanical Engineering from National Taiwan University, and a Ph.D. in Mechanical Engineering from Pennsylvania State University. He is a senior researcher at Langley Research Center under the National Aeronautics and Space Administration (NASA), and an associate fellow of American Institute of Aeronautics and Astronautics. Director General Chau-lyan Chang specializes in computational fluid dynamics and high-performance parallel computing, with the focus on laminar-turbulent transition and space-time CESE (conservation element and solution element) method. He is well known in the field of high-performance computing in Taiwan and abroad for his ability to accurately predict the behavior of high speed, complex fluid properties and flow fields, especially in challenging aerospace applications such as hypersonic flow and turbulence. He has a great reputation in the field of high-performance computing in Taiwan and abroad. The design/engineering software developed by him is widely used not only by NASA but also by U.S. industry, national laboratories, and academia.