Performance summary charts
Here are the summary charts of 35 games and 2 synthetic tests. The highest settings are always chosen and it is DX11 or DX12 with the settings at ultra or maxed except for DOOM which show both Vulkan and OpenGL results (different areas were benchmarked with each API). Specific settings are listed on the main performance chart.
The benches were run at 1920×1080, 2560×1440, and at 3840×2160. All results, except for Firestrike and Time Spy, show average framerates, and higher is always better. In-game settings are fully maxed out and they are identically high or ultra across all platforms. “X” means the benchmark was not run as many new games have been added in the past 6 months to BTR’s benchmarking suite.
The first three columns are devoted to the RX 480 results with the latest drivers 16.12.1 being used in Column 1, and the July 16.7.2 drivers compared in Column 2, with Column 3 devoted to the launch drivers, 16.6.2. The 290X results are given in the next two columns with Column 4 using the latest drivers and Column 5 using 16.6.2 drivers. The last two columns show the GTX 1060 results with Column 6 showing the latest 376.33 drivers and the last Column 7 giving the performance results with the 368.64 launch drivers.
Please open the charts in separate windows or tabs for better viewing.
Our charts provide a lot of information across 35 games and 2 synthetics. What we can take away from the results generally is that the 290X and the RX 480 have gained some significant ground on the GTX 1060 over the past six months.
We were very surprised that the latest NVIDIA driver WHQL 376.33 has seen some serious performance drops with at least 6 games compared with the GTX 1060 launch driver. There are significant performance drops in Rainbow 6 Siege, Shadows of Mordor, Batman AK, Rise of the Tomb Raider, Warhammer, and Fallout 4 using the GTX 1060 6GB compared with the newest driver. There are certainly some performance improvements in the rest of the benchmark suite, but the decreases are surprising.
Of course, it’s always possible that game patches have changed performance since the launch. So we tried these six games back to back with the launch driver versus the new driver just to make sure that something else wasn’t at play. And the launch drivers in each of these six cases outperformed the latest drivers by similar margins to what is reported on the charts.
We also note that AMD has made good performance progress with the RX 480 since its launch, but much of the performance increases already came a month after its launch with 16.7.2. And what is perhaps even more surprising is that the 290X has continued to improve at nearly the same rate at the RX480!
Let’s head for our conclusion.
The GTX 1060 is still the fastest of the three cards that we have tested using our maxed-out 35 game benchmark suite when you consider its overclocking ability. However, the RX 480 and the 290X have managed to gain some real ground on the GTX 1060. We can also conclude that AMD has not neglected Hawaii at all, but each driver revision tends to bring more performance than the last. On the other hand, we were very surprised to see that the GTX 1060 actually lost some performance in a few games with the latest driver compared with its launch driver.
BTR has purchased a Oculus Rift this week, and we have been playing VR games with great success on our GTX 1060. We plan to compare the experience of playing VR games on AMD versus playing on NVIDIA cards in the coming weeks. Stay tuned.
In the meantime, we are currently running a contest that ends on Christmas day. It features a really nice prize bundle for the grand prize winner that is supplied by Kingston with the choice of any PC game under $60 going to the runner up! If you want to enter, join BTR community forums and tell us why you deserve to win. Everyone who enters is a winner as our tech discussions are among the best to be found anywhere!