AMD has just confirmed during its 2015 Q2 earnings call that the Radeon R9 Nano is launching in August.

"Fury just launched, actually this week, and we will be launching Nano in the August timeframe." Said AMD's president and CEO Lisa Su at the 2015 Q2 earnings call.

This is the company's third enthusiast graphics card based on the Fiji GPU featuring 3D high bandwidth memory, or HBM for short and the company's very first reference designed mini-itx form factor enthusiast graphics card.

The R9 Nano purportedly features - although not yet confirmed by AMD - a full fat Fiji XT chip with all resources unlocked, much like the R9 Fury X. However, the differentiator will be the clock speed, which will be much more conservative than the 1050mhz frequency that the company's R9 Fury X flagship boasts. And this is what's going to help the R9 Nano deliver its impressive performance and power efficiency figures.

AMD Radeon R9 Nano Has The Makings Of The Most Powerful Mini-ITX Graphics Card

Based on AMD’s largest, most powerful and latest GPU code named “Fiji”, this tiny card boasts some seriously impressive stats. 2X the performance per watt and 2X the performance density of AMD’s previous flagship, the R9 290X. The result is a six inch long, 175W graphics card that we’re told is “significantly faster” than the 290X. Combine all of those attributes and the R9 Nano has the makings of the the most powerful mini-itx graphics card we have seen so far.

2 of 9

Let’s briefly discuss the technology that enables these radical power savings and size reduction. Stacked high bandwidth memory represents the revolutionary step that has been so badly needed in the evolution of graphics memory standards. The first generation of HBM delivers 300% the bandwidth per watt of GDDR5 memory and reduces the overall board area dedicated to memory by a whopping 94% compared to GDDR5.

2 of 9

The second generation of HBM promises to double the bandwidth by doubling the effective speed of the DRAM data cells from 1Gbps to 2Gbps. While also quadrupling the memory capacity for 4-Hi stacks from 1GB to 4GB and enabling 8-Hi stacks with a whopping 8GB of total memory. AMD has confirmed that HBM2 will be used in the company's upcoming FinFET based GPUs in 2016 which are based on the purportedly code named Arctic Islands graphics architecture flagshipped by the “Greenland” GPU.

The R9 Nano is set feature 16 render back-ends for a total of 64 ROPs as well as 256 texture mapping units. Having fully unlocked Fiji XT silicon as the basis of the card translates to a total of 64 GCN units housing 4096 stream processors. AMD's yet to release the full specifications for the card, however we do have reason believe that AMD will equip the R9 Nano with a Fiji XT GPU, rather than Fiji Pro.

Interestingly, Tom's Hardware managed to improve the power efficiency of the AMD Radeon R9 Fury X by 50% just by reducing the power limit of the card. Surprisingly the only tradeoff was a 10% reduction in the average frames per second. Average FPS dropped from 39.7 to 36 while power consumption dropped by nearly 100W from 267W to 170W. That's 5W below the typical board power for the R9 Nano, so realistically speaking it should prove to be quite easy for AMD to deliver on its R9 Nano promises.

it's quite clear that AMD's Fiji GPU and the GCN architecture in general was designed to run at slightly lower frequencies and as a result would gain a massive chunk of power efficiency through the slightest clock speed reductions. We really can't wait to see what this little card can do. More information is bound to trickle in as we get closer to launch so stay tuned.