SEOUL -- Computing systems are mainly based on the von Neumann architecture, which uses separate processor and memory units to carry out data processing tasks. Processing in memory (PIM) is a next-generation chip that combines logic-based complex processing units within a computer memory.
SK hynix, the second-largest memory chip producer in South Korea, has developed PIM with computing capabilities and a sample of GDDR6-AiM (accelerator in memory) that can be adopted for machine learning, high-performance computing, and big data computation and storage has been developed. Accelerators use processing and computation chips. Graphics DDR (GDDR) is specialized for processing graphics more quickly.
"SK hynix will build a new memory solution ecosystem using GDDR6-AiM, which has its own computing function," SK hynix solution development head Ahn Hyun said in a statement on February 16.
GDDR6-AiM adds computational functions to GDDR6 memory chips, which process data at 16 gigabits, SK hynix said, adding that a combination of GDDR6-AiM with a central processing unit (CPU) or a graphics processing unit (GPU) makes certain computation speed 16 times faster.
GDDR6-AiM, which reduces power consumption, would be combined with artificial intelligence chips in collaboration with Sapeon, an AI chip company spun off from SK Telecom. "We aim to maximize efficiency in data calculation, costs, and energy use by combining technologies from the two companies," Sapeon CEO Ryu Soo-jung was quoted as saying.