SK Hynix has announced that the GDDR6 memory will release in 2018, offing 2x the speed of GDDR5 while consuming 10% lower voltages.
SK Hynix has stated that their new GDDR6 memory will be based on 20nm manufacturing process and offer 16Gbps of memory bandwidth per pin. Which allow GDDR6-based GPU to provide memory bandwidths of 768Gbps compared to 548Gbps of memory bandwidth of current GDDR5X on NVIDIA Titan Xp.
The new GDDR6 will deliver some huge benefits, not just in performance but also in the form of power savings, as fewer chips will be required to maintain the same levels of memory bandwidth and each chip will consume less power and therefore produce less heat.
SK Hynix Inc. (or ‘the Company’, www.skhynix.com) today introduced the world’s fastest 2Znm 8Gb(Gigabit) GDDR6(Graphics DDR6) DRAM. The product operates with an I/O data rate of 16Gbps(Gigabits per second) per pin, which is the industry’s fastest. With a forthcoming high-end graphics card of 384-bit I/Os, this DRAM processes up to 768GB(Gigabytes) of graphics data per second. SK Hynix has been planning to mass produce the product for a client to release high-end graphics card by early 2018 equipped with high-performance GDDR6 DRAMs.
The GDDR6 memory will have to compete against HBM2 in the high-end GPU market. The HBM2 offers lower latency and increased levels of bandwidth and capacity per chip, but HBM2 need to integrate on the GPU itself and therefore difficult to manufacture. Where’s the GDDR6 can be use in the traditional way thus easy to implement and cost less.
Only time will tell whether HBM2 or GDDR6 become dominant in the GPU market moving forward as both memory types certainly have their benefits. With an early 2018 release planned, it is also likely that NVIDIA will be releasing their upcoming Volta- series GPUs with GDDR6 memory, though NVIDIA may go with HBM2 for the top-end card.