AMD Has Taped Out 14nm Greenland – Compute Advantage over NVIDIA and Intel?
In a recent conversation with Forbes, AMD’s head of the graphics division at Radeon Technologies Group (RTG), Raja Koduri, revealed some great details about their 2016 GCN plans. During the brief chat, Koduri confirmed that RTG will launch two brand new GCN GPUs in 2016 as part of their plan to “make Advanced Micro Devices more power and die size competitive.”
That means, we have at least two GPUs coming from AMD, which will feature their next generation GCN architecture along with FinFET process. These new GPUs are believed to be mainstream and high-end members of the Arctic Island GPU family as the chip maker has put a lot of emphasis on “High-Performance” in their previous slides.
Now, a new report has surfaced claiming that AMD has tapped out first prototypes carrying Greenland GPU architecture. If you haven’t looked up lately, Greenland is the alleged codename for the Arctic Islands’ flagship that will cater to performance enthusiasts looking to build an extreme PC system.
The Greenland is expected to pack around 18 billion transistors (up from the 8.9 billion found on the Fiji GPU), connected to 16 and 32GB HBM2 memory, in single and dual-GPU configuration. As it goes from 28nm to 14nm architecture, the big GPU would bring the power down in the 40-50% range – something that AMD CEO Lisa Su confirmed during the last quarterly earnings call.
We are also focused on delivering our next generation GPUs in 2016 which is going to improve performance per watt by two times compared to our current offerings, based on design and architectural enhancements as well as advanced FinFET products process technology.
I can’t say for certain but from the looks of things as they are, AMD seems to be all set to take on NVIDIA Pascal and Intel Xeon Phi. Should it act smart, 2016 could be a landmark year for the company. What do you think? Share your thoughts on the topic in the comments section below.