Synthetic Benchmarks

Before I get into testing, let's take a quick look at where the XFX RX580 8GB GTS stands spec wise. So the stock clocks from AMD are 1257MHz/1340MHz, the Sapphire I tested yesterday was tested at 1411MHz though I found out later it should run at 1450MHz for the limited edition model and the GTS comes in at 1405MHz. So right away we know the Sapphire card has a bit of an advantage but the GTS is close to the 1411MHz of the normal Nitro+ card.

So I started my testing in 3DMark. I tested across three variations of the Fire Strike benchmark and then again with the Time Spy benchmark. Fire Strike is DX11 focused and Time Spy is DX12 focused. The results were about what I would have expected given the clock speeds. The 580 GTS outperformed the heavily overclocked 480 Nitro but came in just a touch behind the 580 Nitro in all four tests. In all of the tests, the GTX 1060 is behind both the 480 and the two 580’s as well.

I then tested in our two Unigine based benchmarks. The Valley Benchmark is older but has more comparison results and is run at 1080p. The new Superposition benchmark is more demanding and was run four times. Twice at 1080p with medium and high detail then at 4k and 8k as well. In Valley, the GTS fall behind for some reason but it held up in the Superposition tests. Of course, all of the GTX 1060’s perform better in the Unigine tests though so the 580’s overall didn’t do as well.

My last set of synthetic benchmarks was in the Catzilla benchmark, I tested at 1080p, 1440p, and 4k to cover a range of resolutions. These results weren’t really any different than the Unigine tests. The 6GB 1060’s were ahead in all three and the Sapphire card with its touch of extra clock speed was ahead in the 1080p and 1440p results with the XFX edging out one point in the 4k test.