Report: Nvidia Pascal Flagship To Use Next-Gen HBM2, Will Launch in 2H 2016
AMD started the new year with a number of announcements: the company unveiled their next-gen Polaris GPU architecture, a new “Wraith” stock CPU cooler as well as launched their Opteron A1100 System-on-Chip (SoC) for Datacenter.
While Nvidia has been quieter than AMD, it seems that Green Team will launch their flagship GPUs based on the Pascal architecture around the same time, in second half of 2016 as AMD’s Polaris will. Moreover, the flagship Pascal GPUs based on the 16nm FinFET process are said to use the next generation High Bandwidth Memory, or HBM2, standard.
That being said, the current HBM2 packages can go up to 16-32 GB VRAM with 1 TB/s+ bandwidth, though the enterprise variants such as Quadro and Tesla models should have even more bandwidth.
The rumor comes just a couple of days after Samsung announced that it has officially commenced mass production of the industry’s first 4GB DRAMs package based on the HBM2 interface. The memory standard was originally developed by SK Hynix in collaboration with AMD, however, they have yet to begin the production.
Nvidia is rumored to have their 16nm Pascal GPU lineup ranging from solutions based on GDDR5X and HBM2 packages but the flagship won’t release until 2H 2016 since that is the time when the higher capacity HBM2 packages go in production.
With AMD set to release Polaris GPUs around the same time, Q3 2016 should see a slew of new products arriving making that time frame pretty exciting for fanboys of both chipmakers.
Source: DigitalTimes Korea