New AI Semiconductor Developed by KAIS Offers 2.4 Times More Cost-effective Solution for ChatGPT
A remarkable breakthrough in ChatGPT technology has caught the attention of the world. KAIST, the Korea Advanced Institute of Science and Technology, has successfully developed a cutting-edge AI semiconductor that provides an impressive 2.4 times more cost-effective solution.
Revolutionary Computing Power for Large Language Models
Leading a team of researchers, Professor Joo-Young Kim from KAIST’s Department of Electrical and Electronic Engineering announced the development of an AI semiconductor specifically designed to accelerate the computation of inferences in large language models, the core of ChatGPT.
The semiconductor, known as the ‘LPU (Latent Processing Unit),’ boasts a highly efficient calculation engine that optimizes memory bandwidth utilization. With its ability to perform all the necessary calculations for high-speed compilation and seamless expandability to multiple accelerators, this AI powerhouse is set to transform the landscape of AI technology.
Impressive Performance Surpasses Industry Standards
The adoption of this LPU-based acceleration tool server enhances performance by up to 50% and achieves a remarkable 2.4 times improvement in price-to-performance compared to industry-leading high-performance GPUs, such as the NVIDIA A100. As a result, it has the potential to replace GPUs in data centers that increasingly rely on AI services.
Prestigious Recognition at International Semiconductor Design Conference
HyperExcel, the new company led by Professor Kim Joo-young, successfully showcased their groundbreaking research at the International Semiconductor Design Automation Conference (DAC) and secured the prestigious award for best presentation in the engineering category. As a globally renowned conference that highlights state-of-the-art semiconductor design technology, DAC attracts leading semiconductor design companies and top universities from around the world.
Advancing AI Semiconductor Technology for Large Language Models
What sets Professor Kim’s team apart is their achievement as pioneers in AI semiconductor technology for large language models. This award signifies global recognition of their solution’s potential to significantly reduce the substantial costs associated with assembling large language models.
With this groundbreaking development, Professor Kim Joo-young of KAIST expressed his determination to capture the global market by revolutionizing large-scale artificial intelligence computations and taking the lead in power technology alongside major tech companies.
By Geun-Nan Lim, Hellotti Reporter
While the world is paying attention to the changes that ChatGPT technology brings, KAIS has developed a core AI semiconductor for ChatGPT that is 2.4 times more cost-effective.
A research team led by Professor Joo-Young Kim from KAIST’s Department of Electrical and Electronic Engineering announced on the 4th that they have developed an AI semiconductor that efficiently accelerates the computation of the inference of the large language model used as the core of ChatGPT .
The ‘LPU (Latent Processing Unit)’ AI semiconductor developed by Professor Kim’s research team accelerates the efficient computation of large language model inferences. It is an AI semiconductor with a calculation engine that can maximize the use of memory bandwidth and perform all the calculations required for high-speed compilation, and is easily expandable to multiple accelerators with its own networking.
This LPU-based acceleration tool server increases performance up to 50% and price / performance 2.4 times compared to a supercomputer based on NVIDIA A100, the industry’s leading high-performance GPU. This is expected to be able to replace high-performance GPUs in data centers, where the demand for productive AI services is increasing rapidly.
This research was carried out by Professor Kim Joo-young’s new company, HyperExcel, and achieved the feat of winning the award for best presentation in the engineering category at the International Semiconductor Design Automation Conference (DAC) held in San Francisco on July 12, US time. .
DAC is a representative conference in the field of international semiconductor design, and is a conference that showcases state-of-the-art semiconductor design technology, particularly in relation to electronic design automation (EDA) and semiconductor design asset technology. conductors (IP). World-renowned semiconductor design companies such as Intel, Nvidia, AMD, Google, Microsoft, Samsung, and TSMC participate in DAC, and many of the world’s top universities such as Harvard University, MIT, and Stanford University are also take part.
Among world-class semiconductor technologies, it is very meaningful that Professor Kim’s team is the only one to receive the award as AI semiconductor technology for large language models. With this award, it was recognized on the world stage as an AI semiconductor solution that can significantly reduce the huge cost required for large language model assembly.
Professor Kim Joo-young of KAIST expressed his aspirations by saying, “We will pioneer the global market with the new processor ‘LPU’ for future large-scale artificial intelligence calculations and preempt power technology of big tech companies.”
Reporter Geun-Nan Lim from Hellotti |
#KAIST #develops #ChatGPT #core #semiconductor #times #costeffective
