To strengthen its capabilities in AI research, National Yang Ming Chiao Tung University (NYCU) has been actively planning to establish an AI high-performance computing platform. NYCU held the inauguration ceremony for this platform on August 6th. It is the first computing platforms in a Taiwanese university to be equipped with the latest NVIDIA DGX systems. The platform will include two NVIDIA DGX H100 and three NVIDIA DGX H200 systems. NYCU is building the system to support the accelerated computing needs of the university’s research teams, enhancing its international competitiveness in AI research.
New AI High-Performance Computing Platform Set to Boost AI Research Capabilities
The inauguration ceremony, which underscored NYCU’s commitments to leveraging accelerated computing to advance AI research, included a donation appreciation segment to thank Jumbo Technology’s Executive Director and alumnus Shun-Tsung Hsu for his NT$10 million donation. Attendees included NVIDIA’s Vice President of Taiwan Eunice Chiu and Jumbo Technology’s general manager Hsiao-Tseng Liu, alongside NYCU’s President Chi-Hung Lin, Vice President Ta-Sung Lee, and Information Technology Center Director Zen-Chung Shih.
Currently, the AI high-performance computing platform is equipped with two DGX H100 systems featuring 16 NVIDIA H100 Tensor Core GPUs and NVIDIA AI Enterprise software. Over the next six months, three DGX H200 systems will be integrated. A high-speed parallel file system will also be introduced, along with NVIDIA Quantum-2 InfiniBand switching, allowing the DGX systems to be integrated in high-performance AI infrastructure and enhancing the scalability and flexibility of multi-GPU operations.
The platform, already equipped with resource scheduling and management software, will efficiently and flexibly provide AI computing resources to various research teams within the university.
AI High-Performance Computing Platform Enhances Research Efficiency, Promotes Energy Conservation, and Strengthens Cybersecurity
The AI high-performance computing platform aims to consolidate and provide necessary accelerated computing resources across the campus, improving the efficiency of computational resource deployment. In conjunction with the university’s green energy data center, the power usage effectiveness (PUE) can reach 1.58, significantly conserving energy and reducing carbon emissions. The maintenance team will continue to strive towards resource sharing and energy-saving goals through the efficiency of accelerated computing, NVIDIA software optimizations, and other methods.
Additionally, in an era where cybersecurity is paramount, protecting computing resources remains a critical task. The team will maintain the university’s existing robust cybersecurity measures, including secure remote access, firewalls, and anomaly monitoring, to safeguard the AI high-performance computing platform.
The successful establishment of this AI high-performance computing platform marks a significant milestone for NYCU. It not only enhances the university’s research capabilities but also positions NYCU as a leader in efficient AI innovation in Taiwan and beyond. The collaborative efforts between the university, industry partners, and alumni underscore the importance of collaborations in advancing technology and research. Looking ahead, NYCU is committed to fostering an environment where cutting-edge AI research can thrive, driving forward scientific discoveries and technological advancements.
