What are the MRAM memories and why are they so important for retaining information?
So-called In-Memory Computing has emerged as one of the promising technologies in the next generation of low-power AI semiconductor chips.
“In-memory computing draws similarity to the brain in the sense that in the brain, computation also occurs within the network of biological memories, or synapses,” explained Dr. Seungchul Jung, author of the research.
It is a process that occurs simultaneously with data storage and data processing in the memory, without the need for the process of exchanging data between two points.
This very operational technology has been promoted by Samsung for some time now, especially through the introduction of MRAM or Magnetoresistive Memory, which has significant improvements in terms of operation speed. First of all, they offer endurance, large-scale production and much lower power consumption than with conventional DRAM.
After a long time of research, Samsung has just given a conclusive and encouraging answer about its MRAM memories, which are said to work effectively in a real-world environment.
The company’s latest news announces that the first functional computing system based on MRAM is fully operational and possible. This marks the creation of a new type of memory that can not only store data, but also process and maintain that data.
MRAM and its advanced data management mode
Basically, in the standard computational architecture, data is stored in memory chips and executed on different processors. This scheme can process a large amount of data stored within the memory network itself, without having to move the data.
The company explains that these new memories can process a large amount of data stored within the memory network itself, without having to access the data on an SSD. Because the processing runs highly parallel, power consumption is substantially reduced.
By introducing MRAM memory (which has already reached commercial-scale production in the semiconductor fabrication system) into the realm of in-memory computing, the frontier of the next-generation low-power AI chip technologies is being extended.
Remember that you can contact us 24 hours a day, seven days a week!