Cooling, Noise, and Power
For the last section, I like to take a look at some of the other important aspects of video cards that sometimes get forgotten about when people are mostly (and justifiably) worried about gaming performance. The first of those tests is to take a look at overall power usage. We already know the GTX 1080 Ti has a 250-watt TDP compared to the 180 TDP of the GTX 1080 so going in we know it's going to pull more. To test it I loop through Valley Benchmark while noting the peak wattage pulled from our test bench using a Kill-A-Watt. This includes the power draw of the whole system and it isn’t exactly a low wattage system so keep that in mind. The GTX 1080 Ti ended up pulling a total of 380 watts at the plug compared to the 283 of the GTX 1080 on the same system. This put the GTX 1080 Ti at the top of the charts for a single card, with just the RX 470 and RX 480’s pulling more when paired up in Crossfire. The 1080 Ti, however, pulls power just like it performs so unlike all of the other power efficient cards Pascal cards we have seen in the past year, you do need to make sure you have the power to push it when you pick one up, especially if you are considering an SLI configuration in the future.
This section has been recently refreshed after our decibel meter died, but it is better off for it with a more accurate meter and a slightly better testing methodology. It does, however, mean that the number of cards tested was low, in fact before this review it was just one card. I did, however, bust out a few other cards to fill in the numbers a little to get a better idea of where the GTX 1080 Ti stands. In my regular testing, I was a little concerned that the card sounded significantly louder than the previous Pascal cards. Well, the numbers did confirm this as well with only the RX480 reference design being louder at full speed and the 1080 Ti being the loudest at 50% fan speed. To see what was going on here I added a new graph as well showing the fan speeds of each of the cards tested when running at 100% fan speed. There was no way removing the DVI port would make the GTX 1080 Ti louder. Well, the fan speeds helped fill in why. The GTX 1080 Ti seems to be using the same fan or at least fan speeds as the GTX 980 Ti where the GTX 1080 and 1070 Founders Editions ran at a lower RPM. The GTX 1080 Ti Founders Edition isn’t going to make you go deaf, but I would at least make sure to keep in mind it's going to make a little additional noise.
For my last test, I wanted to take a look at the cooling performance of the card. With Nvidia talking about cooling improvements and even in their demo their card was running at 66 degrees so I came into my testing with high hopes. After running through heaven benchmark with the stock fan settings it was clear that the card they used was not running at stock fan speeds. My card leveled off at a toasty 86 degrees, putting at the top of my charts. Typically Nvidia designed coolers aim for 80 degrees so I was surprised to see the 1080 Ti run any higher than that really but it was solid at those temperatures. I then wanted to see how good the cooling could be so I cranked the fan up to 100% and ran the same test and it peaked at 63 degrees. This shows that the card is capable of running much cooler and with that, there is room for a little overclocking even with the stock cooling configuration though it will be a balance between noise and fan speeds at that point. It should be especially interesting to see how the aftermarket cards perform, clearly the cranked up card with its 250-watt TDP does put out the heat, keeping it cool will be the biggest challenge.