Key Programs
- Key Programs
- Strengthening AI·ICT Industry Competitiveness
About the High-Performance Computing Resource Utilization Support Program
Strengthens the competitiveness of the AI industry by providing high-performance computing resources (GPU/NPU).
Program Objectives
- Strengthen the competitiveness of the AI industry by providing cloud-based high-performance GPU computing resources, supporting large-scale AI model training and research in advanced technologies such as generative AI and physical AI
- Support the research and development of various AI solutions by providing AI semiconductor–based inference services, while expanding the commercialization and utilization market for domestically developed AI semiconductors and enhancing the competitiveness of the Korean semiconductor industry
Program Details
Implementation Framework and Key Roles
Users are selected through a comprehensive evaluation of factors such as the likelihood of achieving project goals and expected impact. Agreements are then signed with the selected users to ensure smooth resource support.
GPU/NPU resources are provided through resource suppliers, along with technical support to help users effectively utilize the resources.
Users’ resource usage fees are settled and rental fees are paid to the suppliers.
Program results are reviewed through user surveys (July) and performance surveys (November), and user performance analysis and performance management are conducted.
AI training programs are provided to improve resource utilization, and user feedback is collected through meetings to identify areas for program improvement.
About AI Research Computing Support
Contributes to enhancing national AI competitiveness by providing ultra-high-performance GPU resources to leading AI researchers in Korea.
Program Objective
To promote the advancement of innovative AI technologies by identifying providers capable of supplying large-scale computing resources (GPUs) and researchers who require such resources, thereby strengthening national AI competitiveness.
Program Details
- Select private cloud service providers capable of securing and supporting high-performance computing resources based on research GPUs (at least 450 H100 GPUs)
- Establish a real-time monitoring system for resource usage and operate a call center to support researchers
Support the installation and configuration of researchers’ development environments by preparing development environment setups, training, and technical demonstration plans. Conduct pre-testing and provide technical support in an environment similar or identical to the actual research environment provided.
- Program Vision and Goals
Build and operate a large-scale GPU cluster with a capacity of 2 PFLOPS
Provide specialized support including incident response, uninterrupted system management, and security
Support research result presentations and technology exchange meetings
Promote research achievements and support the advancement of research results