Samsung’s archrival is showcasing an HBM3E memory chip that could power Nvidia’s Blackwell B100 AI GPU – with 16 layers, 48 GB and 10.24 Gbps transfer speed, it could be the key to making ChatGPT 6 live
Samsung will showcase numerous new products at the upcoming International Solid-State Circuits Conference (ISSCC), including a high-speed DDR5 memory chip, 280-layer QLC NAND flash memory and the world’s fastest GDDR7.
But while Samsung is sure to attract a lot of attention, it’s not the only game in town as its South Korean rival SK hynix will also unveil its new HBM3E DRAM right after Samsung finishes talking about its 3D NAND flash memory at the High Density Memory and Interfaces session.
HBM3E (High Bandwidth Memory gen 3 Extended) is a breakthrough memory technology that delivers a significant leap in performance and energy efficiency and is designed to meet the escalating demands of high-performance computing, AI and graphics applications.
Nvidia has a choice
HBM3E is the 5th generation of HBM and connects multiple DRAM chips together vertically, significantly increasing data processing speed, capacity and heat dissipation.
According to SK hynix, the new memory chip can process data of up to 1.15 terabytes per second, which is equivalent to processing more than 230 Full HD movies of 5 GB each per second. It also features a 10% improvement in heat dissipation, thanks to the implementation of the advanced Advanced Mass Reflow Molded Underfill technology (or MR-MUF).
SK Hynix sees its HBM3E as the driving force behind AI tech innovation, but it could also power Nvidia’s most powerful GPU yet: the B100 Blackwell AI GPU. Micron has stated that it will not release the next generation of its high-bandwidth memory unit, HBM4, until 2025. This has led to speculation that Nvidia may look for an alternative supplier for the B100 Blackwell.
Although Samsung was considered the most likely candidate for this, SK Hynix is well positioned to step in with its new product with its new ‘Shinebolt’ HBM3E memory.
There’s no official word on this yet, but it probably won’t be long before we find out which of the Korean companies Nvidia chooses.