AMD to nibble the ankles of Nvidia this summer with 14nm FinFET GPUs
Look at me, we're still here, we're still going, still making chips, still relevant
AMD says it will ship graphics chips using its next-generation "Polaris" architecture from mid-2016. Crucially, these processors will use 14nm FinFETs, which means they should have better performance-per-watt figures than today's 28nm GPUs.
Let's be clear: today's announcement is timed to catch the hype building around the CES 2016 conference – the annual tech circle-jerk held in Las Vegas – so don't expect a whole lot of actual detail right now.
What we do know is that Polaris is AMD's 4th generation Graphics Core Next (GCN) architecture, and it will apparently support HDMI 2.0a, DisplayPort 1.3, and 4K H.265 encoding and decoding at 60 frames per second. We assume AMD is going to use GlobalFoundries as its fab. The tech will appear in Radeon products that PC makers are testing out right now, we're told.
A goal of the Polaris design is to fit "console caliber" graphics into thin notebooks and displays. The video below has some AMD bods talking up their gear.
Once you cut through the marketecture, we're told the GPU clock frequency has been upped, the instruction buffer has been increased to keep the cores primed with more work and to fit in larger instructions, and then there's the reliance on FinFETs to reduce leakage current.
AMD's also worked on improving its power gating – so that only relevant parts of the die are powered up during any given operation. Polaris will also bring in a new command processor, display engine, multimedia cores, level two cache, memory controller, geometry processor, and compute units. Presumably, the new GPUs will use AMD's HBM memory.
Meanwhile, rival Nvidia is said to be working on shipping graphics chips built by TSMC using a 16nm FinFET process this year. ®