As the first product to use PIM technology, SK Hynix has developed a sample of GDDR6 AiM (Accelerator in memory). GDDR6 AiM adds computational capabilities to GDDR6 memory chips, which process data at 16Gbps. Using GDDR6 AiM in combination with a CPU or GPU (rather than typical DRAM) can increase the speed of a given computation by up to 16x. GDDR6 AiM is widely used for machine learning, high-performance computing, big data computing, and storage.
Memory chips on conventional computer architectures support only specific data access. And many companies, including SK Hynix, are striving to challenge the traditional concept through innovative applications. With the latest technology, it has finally found a breakthrough solution.
SK Hynix has officially announced the development of its new next-generation memory chips with processing power. This refers to PIM (Processing in Memory) memory chips. However, they have not yet told us all the details of what they will be doing at ISSCC later this month. Expect them to announce that they will be bringing their new GDDR6-AiM (accelerator in memory) memory.
SK Hynix announces GDDR6 AIM memory
Although this announcement won’t be made until the end of the month, we already have some information that the company itself has provided about this memory.
SK Hynix’s new memory adds computational capabilities to GDDR6 memory chips that process data at 16 Gbps. Combining GDDR6-AiM with the CPU or GPU increases computational speed by a factor of 16 compared to traditional DRAM. In addition, the company expects this GDDR6-AiM to be used for machine learning, high-performance computing, and Big Data computation and storage.
This new memory operates at 1.25V, a lower operating voltage than the current product, which is 1.35V, as confirmed by the company itself. In addition, the PIM reduces data transfer to the CPU and GPU, which leads to a significant reduction in power consumption, in this case by up to 80%. This helps the company achieve its carbon reduction goals.
The GDDR6 AiM operates at 1.25V, which is lower than the 1.35V operating voltage of existing products. In addition, the PIM reduces data movement to the CPU and GPU, reducing power consumption by 80%. As a result, this helps SK Hynix meet its commitment to ESG management by reducing the carbon footprint of devices employing the product. Clearly, this helps SK Hynix meet its commitment to reduce equipment carbon emissions (ESG management) by driving the adoption of new technologies.
SK Hynix also plans to launch a technology that combines new memory with AI chips in collaboration with SAPEON Inc. an AI chip company that was recently spun off from SK Telecom.
The company’s solutions development manager, Ahn Hyun, said SK Hynix will build a new ecosystem of storage solutions using GDDR6-AiM.
The company wants to keep the momentum going and make this type of semiconductor memory central to computing and looks ahead to a vision of its use on mobile devices such as smartphones.