Results: Power Consumption And Efficiency
We aren’t going to spent a lot of time in this section since very little changes from the Radeon R9 Fury X. The Nano’s pleasantly low power consumption of 10 to 12W at idle is one of the best results we’ve ever measured for an AMD graphics card in this performance range. Multiple monitors don’t change the story either, similar to what we saw from the Fury X.
|Graphics Card Overall||1W||27W||11W|
Gaming / 3D Operation
Here's where the story gets more interesting. AMD states a typical power consumption of 175W. With a cold card, we actually hit this number, which is great if you’re living in northern Alaska and love to keep your windows open. For everybody else, AMD's specifications prove elusive as soon as a full load, 4K resolution and a warm card come into play. The additional 8W of leakage current doesn't sour us on the Nano's power result, though.
Let’s take a look at the power numbers we measured across the entire benchmark suite and compare them to Nvidia's GTX 970 Mini (OC).
The highest power consumption was reported in Metro Last Light and Thief. We used long scenes for those measurements, not just short sections with uncharacteristically high spikes. Our reasoning will become clear in the stress test section below.
But let’s go back to gaming. We’re puzzled by the bars’ evenness. It’s not a secret that Nvidia’s GPU Boost is restrictive when it comes to hitting power targets. Up until now, AMD seemed fairly contrarian. But these results make it look like AMD is putting its foot down as well. You no longer see bursts of power consumption in certain scenes.
What about the gaming performance to go along with these figures? We just saw that AMD's Radeon R9 Nano consumes more power than the GTX 970 Mini (OC) on average. Putting this into context with frame rates gives those numbers a different spin.
It turns out that the Radeon R9 Nano uses more power, but also provides more performance than the Nvidia GTX 970 Mini (OC). So what happens if we calculate the ratio of the two numbers to get watts per FPS?
No doubt, the resulting graph will cause a lot of head-scratching. The Nano’s efficiency is better in one game, similar in two games and not far behind in all the other games compared to Nvidia's GeForce GTX 970 Mini (OC). The exception is Ashes of the Singularity, and the reason why was covered on the previous page. Here’s the efficiency graph:
There’s simply no comparing this to the Radeon R9 Fury X or AMD’s Hawaii-based graphics cards. AMD is declaring war on power companies, it seems. Power Tune with a real and restrictive limit finally works like it’s supposed to.
For those who like their details, we’re also showing the usual gaming loop measurements. The individual rails’ power consumption numbers, including peaks, can be found right here:
|Graphics Card Overall||33W||437W||186W|
A Stress Test that’s Really Not a Stress Test
This used to be where older Radeon cards were pushed to the point of glowing red-hot. Power Tune put an end to all of that though, so try as you might, you can't use your graphics card as a space heater any longer. First let's look at the curves:
The Radeon R9 Nano’s power consumption is actually slightly lower during the stress test than some of our gaming benchmarks due to a constant load that pushes clock rates down quite a bit.
|Graphics Card Overall||16W||332W||182W|