The world of computing has long faced a fundamental problem: dynamic random access memory (DRAM) is very fast but cannot store data; This means that in the event of a power outage, data will be lost and the data can no longer be saved. On the other hand, there is NAND flash memory, which is non-volatile memory as it can store data without the need for a constant power source. However, since they cannot match the speed of dynamic random access memory (DRAM), today's systems rely on both.
In particular, a group of South Korean scientists has developed a prototype that combines the best technologies in a single memory. This type of memory combines speed and data storage capacity without requiring any power, making it an ideal data storage solution for the world of computing.
A team of researchers from the Korea Advanced Institute of Science and Technology (KAIST) has unveiled a new phase change memory (PCM) storage solution, where previous attempts relied on the use of advanced lithography technologies and techniques. It reduces memory size, but high power consumption and high manufacturing costs pose obstacles for developers.
South Korean scientists have reportedly developed a new method to create a memory that uses fifteen times less power than previous attempts at phase-change memory. At the same time, this approach has reasonable manufacturing costs because it solves two of the biggest problems of phase change memories: It is a real breakthrough in the development of the world's first dynamic volatile random access memory (DRAM).
Intel's Optane memory, based on 3D XPoint technology, is an example of a phase change memory solution designed to provide DRAM-like memory.
This commercial product combines the latency characteristics of DRAM and NAND flash data storage. However, the project between Intel and its partner Micron was not successful, as Intel stopped producing this type of memory in 2021 and ended the project entirely in 2022 due to low demand.
One of the researchers said the team's achievement paves the way for stacked 3D memory systems and neural computing. The problem here is that companies have been talking about phase change memory as an alternative to DRAM for more than a decade. Will we see concrete steps to bring such a solution to the market this time?