Hopefully by now you have watched my main GTX 1060 benchmark video where I threw no less than 25 games at Nvidia’s new mid-range contender. In short the 1060 does indeed deliver GTX 980 like performance making it 12% faster than the RX 480 on average. This advantage is however neutralized by the cost per frame ratio, which also sees the 1060 cost 12% more.
The GTX 1060 does consume quite a bit less power and this resulted in lower operating temperatures for the Nvidia graphics card. The design of the Pascal architecture has also enabled Nvidia to operate at greater frequencies, out of the box the 1060 is clocked at least 37% higher than the RX 480.
What I want to know is through overclocking can the 1060 distance itself further from AMD’s RX 480? We already know that the RX 480 is a pretty rubbish overclocker, at least the reference card is anyway. Overclocks seem to range from 1320MHz up to almost 1400MHz. My card hits a wall at just 1350MHz and this allowed for around 5% more performance.
The Pascal based 1080 and 1070 weren’t amazing overclocks but we have been able to squeeze 15 – 20% more out of the board partner cards, over Nvidia’s stock operating specs. If we can boost the 1060’s performance by at least 10% then this could provide Nvidia’s mid-range GPU with a tangible advantage.
Here’s a quick look at the maximum overclocked operating frequency of the cards we tested.
The focus here will be on the GTX 1060 and RX 480, though out of interest the GTX 980 and GTX 970 have also been overclocked. Please note that although the 1060 Founders Edition card has been overclocked to 1731 MHz for the base and 1933 MHz for the boost, it’s actually operating at around 2.1 GHz in our tests due to the way Nvidia’s GPU Boost 3.0 technology works. For a full list of the test system specs please check the video description and please note only AMD and Nvidia reference cards have been used. All right. Let’s do this thing.
First up we’ll check out how the overclocked 1060 faired in Battlefield 4. Our overclock netted us an 11% gain, which placed it a whopping 38% ahead of the reference RX 480. This was also 13% faster than the overclocked 970, which was 8fps behind on 64. The 1070 was reeled in a little, but the 1060 was still 18% behind.
Next up we’ve got ARMA 3, and here the 1060 was able to add a similar 9% performance to it’s original figures with a new average of 51fps. This made it an even bigger 42% faster than the stock RX 480, and 31% faster than the 480 overclocked. These are pretty significant numbers so far, but I’m not sure we’ll continue to see margins this big in our other games tested.
Star Wars Battlefront
Again, we were able to extract 10% extra performance from the 1060, this time in Star Wars Battlefront. This meant we had a 6fps gain, as well as a more reasonable 23% margin over the RX 480, and 14% once both cards were overclocked. The 1060 was now within 17% of the big brother 1070, but was only 8% faster than the overclocked 970.
Just Cause 3
In Just Cause 4 we saw another 11% gain in performance at 1440p through our overclock, so the average gains are starting to become pretty clear. This time the overclocked 1060 was 20% faster than the 480 and 15% faster than the overclocked 480, while it was 5% faster than the stock 980 now and 19% slower than the 1070.
Far Cry Primal
We tacked on 4fps in Far Cry Primal with our overclock applied, which translated to a 9% gain. These 50 frames were good for a 19% margin over the 480, and this was reduced to 14% once both cards were overclocked. This overclocked 1060 was now 9% faster than the 980, and 16% faster than the overclocked 970.
Tom Clancy’s The Division
Our final game tested was The Division, and here we actually found our smallest gain in performance through overclocking. We were only able to extract an extra 3fps, and at 50fps this translated to just a 6% gain. This meant it was still 19% faster than the 480 and 14% faster than the OC RX 480, while it was 11 faster than the 980… So really, the stock performance was relatively better here compared to the other cards and other games.
I was really impressed with the 1060’s performance in terms of power consumption at the stock clock, and with our overclock applied, this total system power consumption figure only increased from 195 watts to 211. This was an 8% increase in consumption, for a 9% gain in performance, so I’m willing to bet this trade off will be considered well worth it for gamers.
Perhaps more significant is the fact that at 211 watts, it was still using 10% less than the stock clocked RX 480 reference card while pumping out an additional 27% performance on average, so once again I’m thoroughly impressed with the 1060 Founders Edition’s efficiency.
Unsurprisingly, the 1060 runs cool too. In it’s stock config we were running at a cool 64 degrees after 20 minutes of gameplay at 1440p, and once overclocked, this number only increased to 70 degrees. All around impressive stuff and a testament to efficiency of the Pascal architecture.
Prior to any overclocking the GTX 1060 was on average 12% faster than the RX 480. Through overclocking we were able to more than double that margin to 27%, however once we compare the maximum overclocked performance of both the 1060 and 480, we find that Nvidia comes out with a 19% performance advantage.
At almost 20% faster than the RX 480 the GTX 1060 provides gamers who are willing to overclock a real advantage. I should point out though the 1060 overclock on the Founders Edition graphics card only netted us 9% more performance on average, but that’s still a lot better than we were able to achieve with the RX 480 reference card.
Of course the real overclocking battle between these two GPUs will take place once the AMD and Nvidia partner cards become available. I actually have the Palit GTX 1060 Super JetStream on hand so you can check out that video here and I’ll also have video on Gigabyte’s GTX 1060 G1 Gaming shortly as well. So far, there’s no word on custom RX 480 cards however, so we’ll have to continue to wait patiently for those.