Cooling, Noise, and Power
My last round of testing is also one of the most important when you have already decided which GPU you want, in this case, the RTX 2070 SUPER. The performance gap between cards with the same GPU is typically not that large, the overclock does play a role. But as you have seen, even with an overclock it is worth a few FPS at most. The biggest difference between different video cards with the same GPU comes down to cooling performance, the noise that the cooler creates, and how much power the card pulls. For the Gigabyte, RTX 2070 SUPER Gaming OC Gigabyte went with their triple fan Windforce design for cooling and they have always been good for noise levels but hit and miss on cooling depending on how complete the heatsink design is. From what I saw in my initial look at the card the Gaming OC has a solid heatsink so I’m curious to see how it performs.
First though I wanted to check out the power usage. For this, I ran two tests. On both, I monitored the power usage of our test system using a Kill-A-Watt. To put things under load I first uses 3DMark to replicate a gaming load on both the CPU and GPU. I was surprised to see that the Gaming OC actually came in 1 watt below the Founders Edition in this test. The second test where I use AIDA64’s stress test to only load the GPU confirmed this one again, being 3 watts less. Even with a small overclock and an extra fan, gigabyte managed to be slightly more power-efficient!
Next, I wanted to check out how the triple fan Windforce cooler did for noise. Normally smaller fans have to run faster and are nosier but Gigabyte counteracts that with their Windforce design that spins one fan a different direction to cut down on turbulence and wind noise with the fans close to each other. To test this I ran three tests. I tested 50% and 100% fan speeds to see how loud the cooler can be in general. Then my last test was while under load from AIDA64 for a half-hour, once the temperatures leveled off I measured the noise output of the fans. In the first test, I was surprised to see that the triple fan design wasn’t that loud at 100% fan speed. The reason for this did show on the fan RPM chart, Gigabyte doesn’t have the three fans spinning as fast as some other coolers. Turning the fans down to 50% had it down to the second quietest card tested as well. But in the load test, it ended up being the quietest, even coming in below the liquid cooled Vega 64.
My last test was the one I was most curious about, with things running so quiet and not using an ultra tall heatsink and large fans, how would the triple fan design cool the overclocked RTX 2070 SUPER. For this, I ran two different tests. Both times I used AIDA64 to load the GPU and I let that run for over a half-hour until the temperatures leveled off and stopped changing. I did this with the stock fan profile to see what everyone would experience out of the box, and then I did it again with the fans turned up to 100%. The 100% test is just to see how much more headroom is left in the cooler. With the stock fan profile, I was impressed to see the RTX 2070 SUPER Gaming OC down below the Founders Edition with its oversized cooler as well as all of the RTX 2060 SUPERS. In fact, it was running cooler than anything else tested. With the fans turned up it was still low but this time not at the bottom of the chart. It did drop from 61 degrees to 47 degrees which isn’t a bad delta. There is still room in the cooler if you overclock the card more or want it to run cooler, not that I think that is needed given the stock profile results.
While doing my thermal testing with the stock fan profile I also pulled out our thermal camera to take a look at things. I was curious to see how much of a hot spot was up under the card with it partially blowing that direction and to see if there were any warmer areas. Up under the center fan, there was a warmer spot, I was surprised the rest was nice and cool though. As for behind the fans none of the temperatures were that warm at all but down at the end of the card, it was running a few degrees warmer than over top of the GPU itself which tells us that the direct contact heatpipes were doing a good job pulling heat out across the card. The hottest spot on the card was the PCB up on the top edge near the center, but at 128.4F I wouldn’t consider it all that hot at all. There was also a warm area on the backplate near the bottom VRM which has been a trend on most of the RTX cards, but overall I didn’t see any issues with the Flir.