First In Memory Computing chip for GDDR6
SK Hynix in Korea is sampling a memory chip that can process machine learning operations to accelerate AI applications with lower power consumption.
The company is sampling the GDDR6-AiM (Accelerator in memory) that adds computational functions to the next generation GDDR6 memory chips to process data at 16Gbit/s. This is the company’s first chip to use processing in memory (PIM) and is being discussed at the ISSCC conference in the US next week.
GDDR6-AiM chip uses a 1.25V supply, lower than the existing product’s operating voltage of 1.35V. The PIM architecture reduces data movement to the CPU and GPU, reducing power consumption by as much as 80%.
- sureCore develops in-memory computing for edge AI
- 三星使用MRAM AI内存计算
- Cognifiber raises $6m for In-Fibre AI photonic processing
This is the start of an eco-system of products, as SK hynix plans to introduce a technology that combines GDDR6-AiM with AI chips in collaboration with Sapeon, an AI chip company that recently spun off from SK Telecom.
“The use of artificial neural network data has increased rapidly recently, requiring computing technology optimized for computational characteristics,” said Ryu Soo-jung, CEO of Sapeon. “We aim to maximize efficiency in data calculation, costs, and energy use by combining technologies from the two companies.”
“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,” said Ahn Hyun, Head of Solution Development who spearheaded the development of the latest technology.
Related in memory computing articles
- Test chip created for Analog in Memory Computing
- European memory company formed
- UK memory startup raises €1.5m, looks to imec
- In-memory computing provider expands funding round
