The Red Devil RX 570 vs. the EVGA GTX 1060 3GB Overclocking Showdown

Intro

This overclocking showdown is a follow-up to the PowerColor Red Devil RX 570 4GB versus the EVGA GTX 1060 3GB evaluation last week.  Today, we have optimized our overclocks with all performance options set to their highest limits to get the most performance from each card.

At stock, the PowerColor overclocked Red Devil-clocked RX 570 won overall in performance over the stock-clocked EVGA GTX 1060 3GB.  This time, we will overclock the EVGA GTX 1060 3GB and the Red Devil RX 570 each as far as they will go to see where they stand in relation to each other fully overclocked.

The Red Devil RX 570 OC

The $189 PowerColor Red Devil version of the RX 570 4GB is factory overclocked up to its maximum boost speeds of 1320MHz from the reference 1270MHz.  The details of our original manual overclocking may be found here.  We found that as long as the ambient temperatures remain cool and the Power and Temperature limits are maximized, it will not throttle even with the stock fan profile.  After much testing, we settled on +30MHz to the core for a 1360MHz boost with memory clocks +250MHz to 2000MHz.  Unlike with the RX 470, we found that higher memory clocks gained significant performance.  It appears that the Red Devil RX 570 is already running very close to its maximum core clock, and we needed +54mV to achieve another 30MHz for complete stability in all of our tested games.

Overclocking the Gaming Edition of the EVGA  GTX 1060 3GB

We devoted a separate evaluation to the EVGA GTX 1060 3GB which you can read here.  It currently is also priced at $189 but comes with a game bundle choice of For Honor or Ghost Recon Wildlands.  As before, we achieved a final stable overclock of +125 MHz to the core which settled in at or above 1974MHz with GPU Boost 3.0 for the majority of our benching as our room was cool for all of our game benchmarks.  This time, we did not need to adjust the fan profile, but left it on automatic.MAX OC 1060-3Our memory overclock remained at +400MHz for its clock of 4404MHz which greatly contributed to the increased performance.  The fan never became obtrusive as we were able to leave it at stock and the GPU remained cool in the mid-70s C.

Testing Platform

Our testing platform is Windows 10 Home 64-bit, using an Intel Core i7-6700K at 4.00GHz which turbos to 4.6GHz for all cores as set in the ASRock Z7170 motherboard’s BIOS, and 16GB of HyperX DDR4 at 3333MHz. The settings and hardware are identical except for the two cards being tested.

We also feature our newest 2017 games, Ghost Recon Wildlands and Mass Effect: Andromeda, and we also include nine DX12 games including Sniper Elite 4.  We use Unigine’s new Superposition benchmark and we also test VR with Futuremark’s Orange Room benchmark.  We will compare the performance of 26 modern games at 1920×1080 and at 2560×1440 resolutions with maximum settings.

Before we run benchmarks, let’s check out the test configuration.

  • Alexander Yordanov

    Whilst I agree with the testing I strongly oppose getting a gaming GPU with under 4GBs of VRAM in 2017. 2 or 3 GB is good for RX 550/560 and GTX 1050 class GPUs, but not for anything more expensive.

    All this needs is one mod and it will be at “Press F to pay respect” level.

  • cris levin

    They have 570 for $169, or 580 for $209. And you pick a $189 570 to bench? srsly?

    • apoppin

      Yes. It’s what PowerColor sent us for review. It’s a very nice card.

  • aycee

    Thoughts on running this on an R5 build?

    • kaandonmez

      Not too much different I think. (Maybe +/- 3-5 fps most of time) But we all know : in gaming ==>> intel cpus > amd cpus ; but you can buy cheap H110 chipset m.b. and add more money to gpu. So you can c more fps at the same money..