Keynote

Keynote

Presenter:

Dr. Yuan Bo
Innovation Research Institute at Aerospace Legione Technology Co., Ltd.

Title:

Applications and Challenges of High-Performance Computing Accelerated by GPU Clusters in Engineering Simulation

Abstract:

As the demand for computational resources in engineering simulation continues to grow, high-performance computing (HPC) accelerated by GPU clusters has emerged as a pivotal force at the forefront of technology. This report provides an in-depth analysis of the widespread applications, significant advantages, main challenges, and future trends of GPU cluster-accelerated HPC in engineering simulation. By parallel processing vast amounts of simulation data, GPU clusters can significantly enhance the speed and efficiency of engineering simulation tasks, thereby accelerating product design cycles and boosting innovation efficiency. The report begins by introducing the basic architecture of GPU cluster acceleration technology and its support mechanisms for parallel computing. It then discusses how this technology plays a crucial role in fields such as fluid dynamics simulation, structural strength analysis, composite material behavior prediction, and the development of new energy systems. These application cases highlight the unique advantages of GPU clusters in handling complex, large-scale computational tasks. Furthermore, this report explores the challenges faced in implementing GPU cluster-accelerated HPC, including the requirements for advanced programming skills, algorithm optimization, resource management, and scheduling. An assessment of existing solution strategies suggests the use of advanced programming frameworks and the adoption of machine learning technologies to automate and optimize the process. Finally, the report looks forward to the application prospects of GPU cluster-accelerated HPC in future engineering simulations, emphasizing its potential in enhancing computational efficiency, fostering interdisciplinary research, and developing more complex simulation models. With technological advancements and continuous optimization of algorithms, GPU cluster-accelerated high-performance computing is expected to play an increasingly central role in the field of engineering simulation, driving scientific research and technological innovation to new heights.

Biography:

Dr. Yuan Bo (1984~) holds a PhD and focused on research in the field of high-performance computing. He received his Bachelor’s and Master’s degrees in computer science from Beijing Institute of Technology, and a PhD in Computer Science and Technology from Beihang University. After years of experience in the tech industry, he has worked for renowned listed companies both in China and abroad, including IBM, Transwarp and Sinodata, holding key technical positions such as senior programmer, system architect, technical director, and CTO. Yuan has successfully led and participated in the development of several innovative high-performance computing products, earning widespread recognition in the industry for his technical insight and innovation capabilities. With rich practical experience in the fields of Information Technology (IT) and Data Technology (DT), he possesses a profound understanding of industry trends and technological developments.

Currently, he serves as the Director of the Innovation Research Institute at Aerospace Legione Technology Co., Ltd., also holding positions as a part-time director and senior technical advisor at the Beijing Blockchain Technology Application Association, as well as the Deputy Director and Secretary-General of the Computing Power System Simulation Committee of the China Simulation Society.


Keynote

Presenter:

Prof. Gary Tan
Department of Computer Science at the School of Computing, National University of Singapore (NUS)

Title:

A Learnable Behavioural Model for Crisis Management Simulation

Abstract:

Simulation of evacuations in emergencies is crucial in preparing authorities to mitigate disastrous outcomes from unforeseen crises. In an effort to increase the effectiveness of such critical systems, several works have attempted to introduce intelligence in Multi-Agent Systems (MAS) for crisis simulation by incorporating psychological behaviours learned from the social sciences or by using data-driven machine learning models with predictive capabilities. We borrow from Social Sciences behind human characteristics to design a Conscious Movement Model (CMM) to dynamically reflect human behaviour in crisis situations. We then introduce a novel methodology to train this CMM model through pedestrian tracking from video footage using a CMMA (Conscious Movement Memory Attention) model. The trained behaviour model can then be used to inject realism in the crisis simulation. Experimental results on real-life case studies of emergency evacuations show that we can produce realistic simulations similar to actual results, performing better than state of the art methods.

Biography:

Gary Tan is an Associate Professor in the Department of Computer Science at the School of Computing, National University of Singapore (NUS). He received his M.Sc and Ph.D from the University of Manchester, UK.

He is the recipient of numerous Teaching Awards for his excellence in teaching. His research interests include parallel and distributed computing, load balancing, parallel and distributed simulation and High Level Architecture. He is currently working on Digital Twins (Symbiotic Simulation), Crisis Management Simulation and Traffic Simulation.

He is also Vice Dean of Student Life at the School of Computing, and also Master of the College of Alice & Peter Tan, a residential college at NUS.

He has served on several program committees of simulation conferences and symposiums such as AsiaSim conferences and Distributed Simulation and Real-Time Applications (DS-RT). He was program Co-chair of DS-RT2023, held in Singapore. He is a Fellow member of the ASIASIM federation and is also President of the Society of Simulation and Gaming of Singapore. He is on the Editorial Board of the Journal of Simulation Modelling Practice and Theory (Elsevier) and International of Modelling, Simulation and Scientific Computing (World Scientific).