GPU licensing is specifically designed for training and research environments

GPU licensing is specifically designed for training and research environments

Elice

9/21/2023

KAIST

Industry, Regulation

  • University
  • More than 1,000 people

Introduction to servers

  • GPU

Purpose

Research and projects within the National Program for Excellence in SW

KAIST

KAIST is a national specialized university in South Korea that was founded to develop scientific talent and carry out research in science and technology that is supported by national policies. It has consistently developed science and technology talent and spearheaded the advancement of science and engineering in Korea since its founding as the nation’s first research-focused special graduate school for science and engineering. Through active research initiatives including fundamental research and convergence research for science and technology innovation, it generates new values and advances national development.


GPU license for research and projects within the National Program for Excellence in SW

The “National Program of Excellence in Software” is an initiative that helps the Ministry of Science and ICT develop professionals in the field of software engineering by utilizing software to innovate university teaching. KAIST is working to develop inclusive global AI and SW talent after being chosen as a SW-centered university for 2022.
Research and projects including capstone projects, industry-academic collaboration research projects, and start-up promotion programs under the National Program for Excellence in SW require a lot of computer power. Elice was chosen by KAIST to launch the Elice GPU server in order to establish a public cloud license as a credit concept that could adapt to the different GPU cards and performance requirements needed in training and research contexts.


Free GPU credit is offered

On GPUs, “credits” refer to the unrestricted use of resources from the entire amount of GPU computing that has been purchased. The GPU licenses that KAIST introduced are credit-based and can be used up in a way that deducts from the overall sum after a year. The credit technique allows you to allot as many GPU resources as you require and is suitable for activities of your own size as well as learning and research. The same rate structure can be used to cover additional expenses if they arise after the credit has been used up. A billing account can even be created for credit management.


Resources for high-performance computing to address research challenges

The three main categories of KAIST GPU use include general GPU-based research tasks, industry-academic partnership subjects, and high-performance GPU-based research tasks. Given that research objectives necessitated big GPU computing, it was essential to be offered two CPU forms with at least 20 cores. Depending on where they were used, Elice GPUs came with varying delivery times and credits.

Recovery and provisioning of free resources

Elice GPU resources have been restored and made available for unrestricted use by KAIST researchers. The most recent A100-based GPU cards were instantaneously supplied and made available by the user via cloud portal (user console) manipulation. When you provide a system, you make it available for use right away if needed by allocating, deploying, and distributing system resources to suit your needs. Elice GPUs are immediately given to customers via the cloud gateway, scalable based on usage, and designed to promptly restore resources upon the accomplishment of goals and uses. It’s possible to deliver systems that satisfy your requirements and save operating costs with this free resource provisioning and recovery.

Elice provides integrated dashboard features for GPU management by creating a domain specifically for each business and organization. You can monitor GPU quotas and real-time consumption of additional processing resources with dashboards. On the dashboard, you can quickly examine the status and utilization of all assigned instance resources, as shown in the image below. You may view resource use immediately for recently installed NPU and check GPU and CPU utilization, memory utilization, storage, etc. by date. These features let you request specialized GPU instances all at once based on company needs, and they facilitate the efficient use of a particular resource.



▲ The dashboard screen displays the consumption of instance resources.


▲ NPU/GPU On-Demand Dashboard Screen


GPU licenses have been introduced by KAIST with a number of considerations in order to run projects and research. Elice GPU licenses, which satisfy numerous requirements such as high-performance GPUs, domestic data centers, and resource provisioning and recovery, have made it possible for business research and projects under the National Program for Excellence in SW to move forward smoothly.


Human resources, data, and computational resources form the basis of AI research. As a result, the relevance of high-performance computing is becoming increasingly highlighted in research. Elice advises using top-notch GPU resources at a fair price if you need to introduce GPUs for a variety of uses, such as AI education, research, and competitions.


👉 Learn more about Elice GPU

👉 Get the best GPU for your business or school


*Elice owns the copyright to this content, which is protected under copyright law.

*Without prior consent, secondary processing and commercial use of the content are prohibited.

Bring innovative DX solutions to your organization

Sign up for a free trial and a business developer will provide you with a personalized DX solution consultation tailored to your business