엘리스와 함께한 DX 고객 사례
먼저 디지털 전환을 위해 움직이고 있는 기업과 기관은 엘리스의 솔루션을 어떻게 활용하고 있는지 알아보세요.
GPU licenses optimized for research and education
KAIST is a national specialized university in the Republic of Korea established to foster scientific talent and conduct science and technology research promoted by national policy. Since its establishment as Korea’s first research-oriented graduate school of science and technology, KAIST has been leading the development of science and engineering in Korea by continuously cultivating scientific and technological talent. Through its active research activities, including basic research and convergence research for scientific and technological innovation, KAIST creates new value and contributes to the development of the country. GPU license for SW-centered university project research and projects The SW Center University Program is a project of the Ministry of Science and ICT to support the development of SW professionals by transforming university education into a software-centric environment. KAIST was selected as a SW-centered university in 2022 and is working to foster inclusive global AI-SW talent. A large amount of computing resources are required to conduct research and projects such as industry-academia collaborative research projects, capstone projects, and startup promotion programs within the SW Center University project. KAIST wanted to introduce a public cloud license with a credit concept to flexibly accommodate various GPU card and performance specifications required in research and education environments. An announcement to introduce GPU licenses was posted on Nara Marketplace, and Elice was selected among several companies, and three Elice GPU servers were introduced. GPU Credits to spend freely In GPUs, “credit” refers to the right to freely use resources from the total amount of GPU compute purchased. At KAIST, GPU licenses are provided on a credit basis and are used up within a year as a deduction from the total amount. Through the credit system, you can allocate as many GPU resources as you need and use them appropriately for learning, research, and work at the desired scale. In addition, if additional costs are incurred after the credits are exhausted, the same fee system is provided for settlement of additional charges, and you can even set up a billing account for credit management. High-performance computing resources for research projects KAIST’s GPU usage can be broadly categorized into high-performance GPU-based research projects, industry-academia-curriculum collaboration projects, and general GPU-based research projects. In particular, since some of the research projects require large-scale GPU computing, there was a requirement that they must be provided in the form of 2CPUs with 20 or more cores. The number of hours and credits offered for the Elice GPU varied depending on the usage. Provision and reclaim resources freely The Elice GPU resources were provisioned and reclaimed for free use by KAIST researchers. The latest GPU cards based on the A100 were provisioned and provided instantly through user operation on the cloud portal (user console). Provisioning means allocating, placing, and deploying system resources to meet the needs of the user, and having the system ready to go when needed. The Elice GPU was built to be delivered to users instantly through the cloud portal according to the scale of use, and to reclaim resources as soon as the use and purpose are fulfilled. This free provisioning and reclaiming of resources allows the system to be tailored to the user’s needs and reduces operating costs. KAIST has introduced GPU licenses for various aspects of its research and projects. By adopting the Elice GPU license, which meets the requirements of high-performance GPUs, domestic data centers, and resource provisioning and reclamation, we were able to smoothly conduct research and projects for the SW Center University Project. People, data, and computing resources are at the core of AI research, and the importance of high-performance computing in research is becoming increasingly important. By deploying NVIDIA A100-based GPUs, Elice provides a high-performance GPU environment optimized for research and education. GPU resources can be utilized according to the requirements and purposes of each company, and resources of CPU, RAM, GPU, and DISK can be efficiently distributed and allocated. If you need to introduce GPUs for various purposes such as AI education, research, and competitions, we recommend using excellent GPU resources at a reasonable price at Elice. 👉 Learn more about Elice GPU 👉 Get a GPU consultation for your business or school *This content is a copyrighted work protected by copyright law and is copyrighted by Elice. *The content is prohibited from secondary processing and commercial use without prior consent.