이미지 확대
[Courtesy of CJ OliveNetworks]
High-performance computing (HPC) is used to solve extremely complicated computational problems in the science and engineering sectors. It is mainly done by using a supercomputer which plays a key role in the computational science field. The performance of a supercomputer is normally measured in floating-point operations per second (FLOPS). A 1 petaFLOPS (PFLOPS) system can perform one quadrillion FLOPS. The HPC-AI public infrastructure will have a performance capacity of 6 petaFLOPS.
CJ OliveNetworks said in a statement on January 20 that the company won a contract to set up an HPC-based AI public infrastructure at Gwangju Institute of Science and Technology (GIST). An ultra-large computing system that can store 10 petabytes (PBs) will be used by researchers, students and companies to carry out AI-based research projects. A petabyte is a super large unit of digital data that is equal to 1,000 terabytes. A single terabyte can store about 1,000 copies of an encyclopedia.
"Because of the COVID-19 pandemic, the use of HPC is becoming crucial in major countries. For example, advanced computing functions of HPC are being used to develop treatments and vaccines," CJ OliveNetworks official Seo Kyung-ho was quoted as saying.
Copyright ⓒ Aju Press All rights reserved.
View more comments