The PIM technology developed by the Korean firm offers a solution for issues pertaining to data congestion for AI and big data by adding computational functions to semiconductor memory
The first sample of GDDR6-AiM developed by SK hynix which adopts its PIM technology. (Credit: SK hynix Inc.)
South Korea-based SK hynix said that it has developed a next-generation memory chip with computing capabilities called Processing In Memory (PIM).
According to the chipmaker, the PIM technology offers a solution for issues pertaining to data congestion for artificial intelligence (AI) and big data by introducing computational functions to semiconductor memory.
SK hynix revealed intentions to showcase its PIM development during The International Solid-State Circuits Conference (ISSCC) to be held later this month in San Francisco.
The South Korean firm said that continued efforts for innovating the technology to add memory-centric computing capabilities is one step near the reality in devices such as smartphones.
SK hynix has produced a sample of GDDR6-AiM, which is the first product to have adopted the PIM technology.
The GDDR6-AiM is said to add computational functions to Graphics DDR6 (GDDR6) memory chips, which can process data at a speed of 16Gbps.
The company claimed that the combination of GDDR6-AiM with a central processing unit (CPU) or graphics processing unit (GPU) can make certain computation speed 16 times faster when compared to a typical dynamic random access memory (DRAM).
SK hynix said that GDDR6-AiM is largely anticipated to be adopted for high-performance computing, machine learning, and computation and storage of big data.
GDDR6-AiM is claimed to run on 1.25V, which is lower than the operating voltage of 1.35V of the existing product. Additionally, the PIM technology cuts down data movement to the CPU and GPU, thereby decreasing consumption of power by 80%.
SK Hynix solution development head Ahn Hyun said: “SK hynix will build a new memory solution ecosystem using GDDR6-AiM, which has its own computing function.
“The company will continue to evolve its business model and the direction for technology development.”
SK hynix also intends to launch a new technology for combining GDDR6-AiM with AI chips in partnership with Sapeon, an AI chips developer that spun off recently from SK Telecom.
Sapeon CEO Ryu Soo-jung said: “The use of artificial neural network data has increased rapidly recently, requiring computing technology optimised for computational characteristics.
“We aim to maximise efficiency in data calculation, costs, and energy use by combining technologies from the two companies.”
Last month, SK Hynix alongside SK Telecom and SK Square set up a fund of around $835.1m to invest in firms with innovative technologies like AI, metaverse, blockchain, and semiconductors.