Today we are testing the newest high-end video card from AMD, the 8 GiB Radeon RX 580. Check it out!
The Radeon RX 580 is based on the Ellesmere chip, which is the same used on the Radeon RX 480. Its is manufactures under 14 nm process with 3D “FinFET” transistors.
Like the Radeon RX 480, the Radeon RX 580 GPU has 2,304 cores, divided in 36 Compute Units, and uses a 256 bit memory bus. Their standard clocks, however, are higher: 1,257 MHz base clock and 1,340 MHz boost clock. Its TDP is 185 W. There are 4 GiB and 8 GiB versions, both with memory running at 8 GHz.
According to AMD, it supports FreeSync, Vulkan, VR Premium, and DirectX 12 technologies. Besides that, this new generation comes with the “Radeon Chill” feature, that uses algorithms on the drivers to reduce the power consumption without reducing the performance.
The model we tested is the XFX RX-580P8DFD. It brings higher clocks than the reference model: 1,366 MHz base clock, 1,386 turbo clock, and memory working at 8,1 GHz on boost mode.
Figure 1 shows the tested XFX Radeon RX 580. The cooler frame is made of plastic with a finishing that looks like carbon fiber.
Figure 1: the GeForce GTX 1070 Founders Edition
Costing from USD 230 (on the simplest versions,) the Radeon RX 580 hits the market on the same price point of its predecessor, the Radeon RX 480, and costing the same of its direct competitor, the GeForce GTX 1060 6 GiB. So, we benchmarked the Radeon RX 580 against those two video cards.
In the table below, we compare the main specs from the video cards we included in this review. Prices were researched at Newegg.com for this article.
Video card |
Core clock |
Turbo clock |
Effective memory clock |
Memory bus |
Memory bandwidth |
Memory |
Processing cores |
TDP |
DirectX |
Price |
Radeon RX 580 |
1,366 MHz |
1,386 MHz |
8.0 GHz |
256 bit |
256 GB/s |
8 GiB GDDR5 |
2,304 |
185 W |
12 |
USD 280 |
GeForce GTX 1060 FE 6 GB |
1,506 MHz |
1,708 MHz |
8.0 GHz |
192 bit |
192 GB/s |
6 GiB GDDR5 |
1,280 |
120 W |
12.1 |
USD 230 |
Radeon RX 480 |
1,120 MHz |
1,266 MHz |
8.0 GHz |
256 bit |
256 GB/s |
8 GiB GDDR5 |
2,304 |
150 W |
12 |
USD 230 |
Now let’s take a closer look to the tested video card.
[nextpage title=”The GeForce GTX 1070″]
The XFX Radeon RX 580 is 10.9 inches (278 mm) long and uses two slots. It uses two 100 mm fans. An amazing detail is that its fans are turned off while the GPU is idle.
Figure 2 shows the video connectors of the Radeon RX 580. It brings three DisplayPort 1.4, one HDMI and one DVI-D connectors.
Figure 2: video connectors
In Figure 3, you see the top of the card, where there is an eight-pin PCI Express power connector.
Figure 3: top view
Figure 4 shows the bottom of the card.
Figure 4: rear view
Figure 5 shows the solder side of the RX 580, which is covered by a backplate.
Figure 5: backplate
In Figure 6 you see the Radeon RX 580 with the backplate removed, unveiling the solder side of the PCB. There are no memory chips on this side.
Figure 6: solder side
Figure 7 shows the XFX Radeon RX 580 with its cooler removed. It is big part cooler, with four heatpipes, that also cools the memory chips and the voltage regulator circuit transistors. Notice that there are eight memory chips.
Figure 7: cooler removed
Figure 8 unveils the Ellesmere chip, which is manufactured under 14 nm process.
Figure 8: the GP104 chip
Figure 9 shows one of the memory chips present on the video card. It is a GDDR5 Samsung K4G80325FB-HC25 memory chip, whose nominal clock is 8 GHz.
Figure 9: memory chip
Figure 10 shows the voltage regulator of the XFX Radeon RX 580. It uses six phases for the GPU and one phase for the memory chips.
Figure 10: voltage regulator
[nextpage title=”How We Tested”]
During our benchmarking sessions, we used the configuration listed below. Between the tests, the only variable component was the video card being tested.
Hardware Configuration
- CPU: Core i7-6950X running at 3.8 GHz
- Motherboard: ASRock Fatal1ty X99 Extreme6/3.1
- CPU Cooler: Thermaltake Water 3.0 Ultimate
- Memory: 64 GiB DDR4-3000 HyperX Predator, four KHX3000C15/16GX 16 GiB modules running at 2400 MHz
- Boot drive: Kingston HyperX Savage 480 GB
- Case: Thermaltake Core P3
- Video Monitor: Samsung U28D590
- Power Supply: Corsair CX750
Operating System Configuration
- Windows 10 Home 64-bit
Driver Versions
- AMD driver version: Crimson 16.12.1
- NVIDIA driver version: 376.33
Software Used
- 3DMark
- Deus Ex: Mankind Divided
- Dirt Rally
- Gears of War 4
- Hitman
- Mad Max
- Rise of the Tomb Rider
- Tom Clancy’s Rainbox Six Siege
Error Margin
We adopted a 3% error margin; thus, differences below 3% cannot be considered relevant. In other words, products with a performance difference below 3% should be considered as having similar performance.
[nextpage title=”3DMark”]
3DMark is a program with a set of several 3D benchmarks. We ran the Time Spy, Fire Strike Extreme and Sky Diver benchmarks.
The Time Spy benchmark measures DirectX 12 native performance, running at 2560 x 1440 resolution. In this benchmark, the Radeon RX 580 8 GiB was in a technical tie with the GeForce GTX 1060 6 GiB, and was 7% faster than the Radeon RX 480 8 GiB.
The Fire Strike Extreme benchmark measures DirectX 11 performance and is aimed at high-end gaming computers. It runs at 2560 x 1440. In this test, the Radeon RX 580 8 GiB performed similarly to the GeForce GTX 1060 6 GiB, and was 9% faster than the Radeon RX 480 8 GiB.
The 3DMark Sky Diver benchmark is aimed at mainstream PCs with DirecX 11 simulations. It runs at 1920 x 1070. In this test, the Radeon RX 580 8 GiB also was in a technical tie with the GeForce GTX 1060 6 GiB, and was 7% faster than the Radeon RX 480 8 GiB.
[nextpage title=”Gaming Performance – Part 1″]
Deus Ex: Mankind Divided
Deus Ex: Mankind Divided is an action RPG with FPS elements, launched in August 2016, that uses the Dawn engine, being compatible with DirectX 12. We tested it using the benchmark included in the game, with DirectX 12 enabled, graphic options “very high” and MSAA 4x.
The results below, in Full HD and 4K, are expressed in frames per second.
On Deus Ex: Mankind Divided in Full HD, the Radeon RX 580 8 GiB was 15% faster than the GeForce GTX 1060 6 GiB, and 4% faster than the Radeon RX 480 8 GiB.
On 4K, the Radeon RX 580 8 GiB was 17% faster than the GeForce GTX 1060 6 GiB, and 8% faster than the Radeon RX 480 8 GiB.
Dirt Rally
Dirt Rally is an off-road racing game launched in April 2015, using the Ego engine. We tested the performance at this game using the built-in benchmark, with quality option as “high” and 2x MSAA.
The results below are expressed in frames per second.
On the Full HD benchmark, the Radeon RX 580 8 GiB was 16% faster than the GeForce GTX 1060 6 GiB, and 11% faster than the Radeon RX 480 8 GiB.
On 4K, the Radeon RX 580 8 GiB was 11% slower than the GeForce GTX 1060 6 GiB, and 7% faster than the Radeon RX 480 8 GiB.
Gears of War 4
Gears of War 4 is a third-person shooter launched in October 2016, using the Unreal Engine 4. In order to measure the performance on this game, we ran the included benchmark, with overall quality set as “high”.
The results below are expressed in frames per second.
On GTA V, with Full HD, the Radeon RX 580 8 GiB was in a technical tie with the GeForce GTX 1060 6 GiB, and was 8% faster than the Radeon RX 480 8 GiB.
With 4K resolution, the Radeon RX 580 8 GiB was 6% faster than the GeForce GTX 1060 6 GiB, and 9% faster than the Radeon RX 480 8 GiB.
[nextpage title=”Gaming Performance – Part 2″]
Hitman
Hitman is an action/stealth game, launched in March 2016, that uses a DirectX 12 compatible version of the Glacier 2 engine. To measure performance in this game, we ran the benchmark in it, measuring the framerate with FRAPS. We ran this game with DirectX 12 enabled, with image quality set as “high” and SMAA on.
The results below, in Full HD and 4K, are expressed in frames per second.
On Hitman, in Full HD, the Radeon RX 580 8 GiB performed similarly to the GeForce GTX 1060 6 GiB, and was 6% faster than the Radeon RX 480 8 GiB.
With 4K resolution, the Radeon RX 580 8 GiB was in a technical tie with the GeForce GTX 1060 6 GiB, and was 6% faster than the Radeon RX 480 8 GiB.
Mad Max
Mad Max is an open-world action game launched in September of 2015, using the Avalanche engine. In order to measure the performance using this game, we ran its intro, measuring the framerate with FRAPS three times. We ran the game with image quality set as “very high”.
The results below are expressed in fps and they are the mean between the three collected results.
On Mad Max, in Full HD, the Radeon RX 580 8 GiB was 5% faster than the GeForce GTX 1060 6 GiB, and performed similarly to the Radeon RX 480 8 GiB.
On 4K, the Radeon RX 580 8 GiB was 19% slower than the GeForce GTX 1060 6 GiB, and was 36% slower than the Radeon RX 480 8 GiB.
Tom Clancy’s Rainbow Six Siege
Tom Clancy’s Rainbow Six Siege is a tactical FPS launched in December 2015, based on the AnvilNext engine, which is DirectX11. In order to measure the performance using this game, we ran the benchmark included on it, using graphics quality as “high”.
The results below are expressed in frames per second.
On this game, in Full HD, the Radeon RX 580 8 GiB was 12% faster than the GeForce GTX 1060 6 GiB, and 6% faster than the Radeon RX 480 8 GiB.
On 4K resolution, the Radeon RX 580 8 GiB was 15% faster than the GeForce GTX 1060 6 GiB, and 6% faster than the Radeon RX 480 8 GiB.
Rise of the Tomb Rider
Rise of the Tomb Rider is an adventure/action game launched in January of 2016, based on Foundation engine. In order to measure the performance using this game, we ran the benchmark included on it, using graphics quality “high” and DirectX 12 enabled.
The results below are expressed in frames per second.
On Rise of the Tomb Rider, in Full HD, the Radeon RX 580 8 GiB was in a technical tie with the GeForce GTX 1060 6 GiB, and was 8% faster than the Radeon RX 480 8 GiB.
On 4K resolution, the Radeon RX 580 8 GiB was 6% faster than both the GeForce GTX 1060 6 GiB and the Radeon RX 480 8 GiB.
[nextpage title=”Conclusions”]
There was a certain disappointment on the launch of the Radeon RX 500 series, because they use the same architecture of the RX 400, only with slightly higher clocks, while the enthusiasts were expecting for AMD’s new architecture, Vega.
This new series is a minor evolution on its predecessor, which is very common in the semicondutor industry, when a technology gets more mature and the chips become more efficient.
That’s why there was no surprise on our tests: the Radeon RX 580 was a little faster than the Radeon RX 480 on most benchmarks. Compared to its direct competitor, the GeForce GTX 1060 6 GiB, the Radeon RX 580 was, on average, similar, being a little faster on some games and a little slower in other ones.
So, we can say the Radeon RX 580 came to replace the RX 480 and fight the GeForce GTX 1060 6 GiB, being a video card with excellent cost/benefit ratio.
Leave a Reply