Meet the 4K AMD Radeon R9 Nano GPU: Launching today and totally built for tight spaces and serious processing needs
Stephan Jukic – August 28, 2015
As of August 27th, AMD has officially launched the very latest and possibly most unique of its existing crop of 4K gaming GPUs. We are of course talking about the AMD Radeon R9 Nano graphics card and what largely makes it so unique is the fact that this very powerful, very beefy GPU is also tiny enough to fit into some very tight little computing spaces.
The Nano is aimed particularly at users who want to have a smaller PC with a mini-ITX motherboard and weaker-sided power supplies but who still want a GPU that does the absolute most it can to deliver heavy-duty power without overloading the power unit or ramping up the heat within its confined space.
This is where the Nano definitely comes into its own and its entire design is aimed at the above needs. For users who don’t care about space, power drain or heat generation too much, then the Nano is not the ideal choice. Instead, for those users, the best option on the current roster would be something like the Radeon R9 Fury X or the massively powerful dual GPU Radeon R9 295×2.
But going back to the Nano itself, this little monster is certainly underclocked but nowhere near enough to be called a weak GPU by any means. On the contrary, its price alone should be a firm indicator of solid performance benchmarks. At $649, the R9 Nano cost pretty much the same as AMD’s Fury X. Furthermore, it features AMD’s AMD’s latest graphics processing chip architecture, the highly acclaimed “Fiji” design.
Fiji, which has been here only since June of this year, is the chip architecture being used in a number of the latest 4K and high frame rate HD GPUs from the company. What makes Fii particularly powerful is the fact that its 28nm silicon design uses what is called High Bandwidth Memory (HBM) by AMD, which is a stacking design technology for much more space-efficient construction. Essentially the DRAM of the Fiji chip is stacked right next to the GPU cores right on the same chip.
With this space efficiency, HBM memory lets the graphics processor access data much faster than is the case with an external GDDR5 memory, which in turn create more bandwidth per watt consumers in the Fiji chips. This is one of the Nano’s main strengths, and what allows it to work with a relatively high level of power while being so small.
Overall, the Nano can perform 8.19 trillion single-precision floating-point calculations per seconf with 64 compute units and 4,096 stream processors. It also offers 4GB of the above-described HBM along with a 4096-bit bus that works at 512GB/s and also offer support for DirectX 12, Vulkan, OpenGL 4.5 and Mantle. In terms of connectivity, the R9 Nano also includes three DisplayPort 1.2 ports and a single HDMI 2.0 port for also gaining graphics output at 4K and 60 frames per second, which is a major innovation which was up until recently the exclusive domain of Nvidia’s latest 4K GPU’s.
Finally, for the sake of this brief overview, the Nano can run its engine clock at up to 1GHz with a more regular speed range of 850 to 950MHz during gameplay.
Overall, the Fiji isn’t quite a true 4K-ready graphics processing unit but it comes very close to being so and if you want to do 1440p gaming at solid frame rates or HD gaming at hardcore frame rates, the R9 Nano is a definite candidate for pulling this off inside a smaller PC space.
Story by 4k.com