Cooling, Noise, and Power
Asus spent a lot of time tuning and tweaking the GTX 670 DirectCU Mini’s design in order to be able to make it all work in such a small form factor. Because of that I was very interested to see what its cooling performance was going to be when I put it on the test bench. I didn’t expect it to outperform all of the cards that are nearly twice its size but I was really hoping to see it out perform the reference design. At 77 degree under load it just that, barely. It’s not a number to be all that proud of when taken on its own, but being such a small card with a high end GPU under the hood I think we can be a little forgiving, not to mention the GTX 670 DirectCU Mini is actually overclocked over the reference design as well.
With the unique fan design I was curious to see how it would perform in my noise testing. I ran the card through both idle and full fan speed tests to see the range of noise that you can expect. At idle the GTX 670 DirectCU Mini came in at the middle of the pack, but the differences between cards are extremely small. Under 100% fan speed the GTX 670 DirectCU Mini did come in a little higher at 75.2 decibels, slightly higher than the reference GTX 670 performed. I can only assume that Asus had to give up some of its noise performance to make the small card possible.
In my power testing at idle the GTX 670 DirectCU Mini was three watts higher than the reference GTX 670. The different cooling design could have played a role in that, along with the higher base clock speed (928MHz vs. 915MHz). Under load the difference was more with the fully loaded test bench with water cooling and a 3860X the GTX 670 DirectCU Mini pulled 403 against the reference cards 378. That difference is without a doubt due to the overclock that Asus included bumping the boost clock up from 980MHz to 1006MHz.