XFX Radeon 6970 and 6950
Temperatures and Power Consumption
To test temperatures, we fire up FurMark and let it run for awhile until the temperatures level out. FurMark pushes GPU thermal loads extremely high, and can be considered a worst-case scenario. We’ve had to slightly tweak our testing methodology since the GTX 500 cards employ a "limiter" that throttle back temperatures under extreme load such as FurMark. Consequently, we include results with, and without, the limiter engaged.
Below are the FurMark "worst case" results:
Well, we have a bit of a mixed bag of results here. The XFX 6950 performs respectably, with a moderate delta and a final load temperature of 81°C, which isn’t terribly impressive or disappointing in either regard. However, the 6970’s load temperature is quite a shocker to be honest, finishing at a very toasty 91°C. Frankly, we were expecting better results, particularly due to the heatsink’s vapor chamber technology being used, as we’ve typically seen this to be an extremely effective cooling solution. The 6970’s main competition, the GTX 570, runs considerably cooler, particularly with the limiter engaged.
We create a baseline power draw without a graphics card installed. After a graphics card is installed, we let the system sit idle in Window for 10 minutes and then run FurMark for maximum power draw. The difference gives us the power consumption for the card. The results below are for the graphics card only. As with the temperature testing, we report results for the Nvidia 500 cards with, and without, the limiter engaged.
Here we see the XFX cards faring better in relation to the competition, as the 6950 sits comfortably low on the chart, even more frugal than the mid-range GTX 460. The 6970 sits on par with the GTX 570 when its limiter is not artificially engaged. Power consumption seems to scale fairly equally among the Radeon cards in proportion to the graphics horsepower available.