In recent years, AMD has been updating its graphics processors quite rarely, moreover, often making a banal name change rather than improving technical characteristics. We have already encountered this more than once; this also happened to the HD7970 video cards, which were later renamed R9 280. And almost the same thing happened to the RX 480, which later became the RX580 with slightly increased frequencies. It turns out that the RX580 GPU is already several years old, but is it still relevant now? We will deal with this issue using the example of the ASUS DUAL Radeon RX 580 8G OC Edition video card.
The ASUS DUAL Radeon RX 580 8G OC Edition video card, although it does not belong to any of the ASUS gaming lines, still boasts quite powerful non-reference cooling and a redesigned circuit board. At the same time, it also costs a little less than its “gaming” counterparts.
According to Yandex.Market, the ASUS DUAL Radeon RX 580 8G OC Edition video card can be purchased for 19,000-20,000 rubles. (Price may vary depending on your region and date of purchase).
Specifications ASUS DUAL Radeon RX 580 8G OC Edition
Packaging and equipment ASUS DUAL Radeon RX 580 8G OC Edition
The ASUS DUAL Radeon RX 580 8G OC Edition video card comes in a standard cardboard box, decorated in dark colors. On the front we see an image of a “predatory cat”, which has become the face of the DUAL line. Also here you can see the name of the video card, information that this is an overclocked version (OC Edition) and a mention of the fully automated production of this model.
On the opposite side, the manufacturer described the main features of the ASUS DUAL Radeon RX 580 8G OC Edition video card:
- use of dust-proof fans with IP5X certification;
- silent operation of the video card when the chip heats up to less than 55 degrees;
- fully automated production process and high quality Super Alloy Power II components.
In the package we found an instruction manual and a disk with drivers and proprietary software.
Appearance and device ASUS DUAL Radeon RX 580 8G OC Edition
The appearance of the ASUS DUAL Radeon RX 580 8G OC Edition video card may seem familiar to many. Indeed, the design of the cooling system casing is very similar to the DirecCU II cooler we previously encountered. Although in essence this is what it is, only with minor modifications. DirectCU II was released several years ago and was so successful that it is still used today. The ASUS DUAL Radeon RX 580 8G OC Edition cooled by it shows very good results, or rather low operating temperatures.
The video card is quite compact by today's standards, its dimensions are 242 x 128.9 x 38 mm. The entire front side is covered with a cooler with a pair of 97 mm fans. By the way, ASUS DUAL Radeon RX 580 8G OC Edition is one of the few models that is not equipped with LED backlighting. So if you really don’t like backlighting, then this model is your choice!
The back side of the printed circuit board is not covered with anything, there is no backplate here, the only thing you can notice is a small aluminum plate that acts as a stiffener that will not allow the video card to sag over time.
ASUS DUAL Radeon RX 580 8G OC Edition is built on high-quality Super Alloy Power II electronic components, which ensure stable operation even when overclocked. Looking ahead a little, our sample was able to run stably at 1450 MHz, which is a very good result for the AMD Polaris 20 XTX (Ellesmere) GPU.
View of the ASUS DUAL Radeon RX 580 8G OC Edition video card from different angles. It takes up only two expansion slots, and if you are planning to build a PC with a pair of such accelerators, then you can do this even in a small MicroATX case.
The interface panel provides the following video outputs:
- 1 x DVI-D;
- 2 x HDMI 2.0;
- 2 x Display Port.
Additional power for video cards is supplied through the 8-pin PCI-E connector. At peak loads, the video card can consume up to 180 W of electricity, which leaves us with a good margin for further overclocking (75 W + 150 W).
The stiffener is secured to the video card using four screws.
The ASUS DUAL Radeon RX 580 8G OC Edition graphics card PCB is a benchmark for many manufacturers. It is made so carefully that you can admire it for hours. And all the credit goes to ASUS proprietary technology – Auto-Extreme. The board itself is made in a non-reference design, the arrangement of all elements is standard.
Around the AMD Polaris 20 XTX (Ellesmere) graphics processor there are eight chips with a total capacity of 8192 MB.
GPU technical characteristics: 2304 stream processors, 144 TMU units, 32 ROP units, 256-bit data bus, operating frequency 1360 MHz.
The memory chips are manufactured by Samsung and are labeled K4G80325FB-HC25. Nominal frequency 2000 MHz. In general, these microcircuits are among the best, they overclock well and are easy to flash with low timings. So if you decide to buy an ASUS DUAL Radeon RX 580 8G OC Edition video card for mining, then you won’t go wrong!
The video card power supply system is made according to a 3+1+1 phase scheme. Three of them are responsible for the GPU, and these phases are doubled using IR3598 drivers, giving us a total of six phases. One phase is dedicated to powering the memory and another is responsible for the PLL.
A compact aluminum radiator removes heat from the heating VRM chips.
M3056M and M3054M microcircuits are used as power assemblies. the first can withstand loads up to 107 A, and the second up to 95 A.
The power converter is controlled by a digital PWM controller VRM DIGI+ ASP1300.
As mentioned above, the video card is cooled by a good, time-tested DirecCU II cooler. It consists of an aluminum fan with a pair of 8mm heatpipes and a pair of 97mm fans. It is secured to the PCB with just four screws.
The heat pipes are in direct contact with the GPU, which is clearly visible in the second photo.
The radiator is made of aluminum plates 0.2 mm thick, it does its job perfectly. True, the memory chips do not contact it; they are content with blowing the fans.
The fans are manufactured by FirstD; their rotation speed can vary from 1150 rpm to 2650 rpm. The fans are very quiet; they begin to make noise only after reaching 80% of maximum speed.
Testing ASUS DUAL Radeon RX 580 8G OC Edition
Test stand:
- Intel Core i5-8600K processor
- Corsair Vengeance LPX DDR4-2800 RAM
- ASUS ROG Maximus X Apex motherboard
- Corsair AX1200i power supply.
Testing of the video card is divided into two parts: first, the video card was tested at nominal frequencies, then the same tests were run at higher frequencies. The graphics processor was overclocked from 1360 MHz to 1450 MHz, and the GDDR5 memory was overclocked from 8000 MHz to 9000 MHz. When there is no load, the video card resets the GPU frequency to 300 MHz, and the memory slows down to 300 MHz.
The test results are presented in the graphs below. The games were set to maximum settings and a resolution of 1920x1080 pixels.
Benchmark testing.
Rated frequencies, fan operating mode - Auto.
High frequencies, fan operating mode - 80% of maximum speed.
Conclusion
The ASUS DUAL Radeon RX 580 8G OC Edition video card turned out to be very successful, it is perfectly balanced and can claim the best quality/price ratio. Finding similar equipped solutions on the market for the same amount is quite problematic.
It’s impossible to get to the bottom of the ASUS DUAL Radeon RX 580 8G OC Edition and say that this is bad, we didn’t find any disadvantages. Yes, the cooler used here is not the most modern, but believe me, it’s enough for the eyes. The video card does not overheat and does not make noise, but what else do you need? By the way, all “non-fans” of RGB backlights can already line up for the ASUS DUAL Radeon RX 580 8G OC Edition, since it does not have backlighting at all. But there is excellent memory on Samsung chips, which overclocks well and flashes with lower timings.
As for performance, it is still sufficient for Full HD resolution, games run at a comfortable FPS, and most likely this will continue for at least another year.
Based on the testing results, we recommend the ASUS DUAL Radeon RX 580 8G OC Edition video card for purchase, and award it the “Wise Choice” award!
Right away AMD Radeon RX 580 Sapphire's overclocked Nitro+ Limited Edition shows an 8% advantage over our overclocked Radeon RX 480. That's enough to rival even the Nvidia GeForce GTX 1060 6GB (also overclocked) in terms of average frame rates. All our measurements confirm that both the RX 580 and GTX 1060 6GB provide equally smooth images.
The draw remains the same at a resolution of 2560 x 1440 pixels. This may come as a surprise to many, especially if you remember our review of the Nvidia GeForce GTX 1060 6GB (http://www..html) last year, in which the Radeon RX 480 8GB was faster than the 1060 6GB. However, at the beginning of this year, Nvidia representatives were already talking about working to improve the performance of their cards in games based on DirectX 12. One of the beneficiaries of this work was the game Ashes of the Singularity, and we are now seeing the results of these efforts in the Escalation version.
Battlefield 1 (DX12)
Radeon RX 580 cements the RX 480's superiority over the GeForce GTX 1060 6GB in Battlefield 1, delivering approximately 9% higher frame rates. In fact, the new Radeon RX 570 leaves Nvidia's card far behind.
Despite the slow start, which affects our minimum framerate measurements, Sapphire Nitro+ Radeon RX 580 turned out to be 15% faster than the Nvidia GeForce GTX 1060 6GB in Battlefield 1 with Ultra quality settings at a resolution of 2560 x 1440. Both cards provide comfortable gameplay, but it seems that Nvidia should reconsider the price in order to maintain the competitiveness of accelerators on the GP106 chip.
Civilization VI (DX12)
Civilization VI is not the most demanding game in our set, but we can still set the maximum requirements and observe some semblance of scalability depending on GPU performance.
I wonder how Radeon RX 580, and the RX 570 was ahead of the RX 480 and 470 in terms of average frame rates. They also show a noticeably higher minimum framerate. However, at the same time, they are inferior to the three Nvidia cards, which we did not expect to see among the leaders (especially the GTX 1060 3GB).
Although Radeon RX 580 outperformed the GeForce GTX 1060 3GB in the average frame rate test at a resolution of 2560 x 1440; it was inferior to the 6GB version and the GeForce GTC 970.
Frame rate or frame time graphs don't really explain why Civilization VI favors GeForce cards so much. Fortunately for AMD, their advantage is generally insignificant. The frame rate at QHD resolution is decent, especially considering the genre of this game.
Doom (Vulkan)
Radeon RX 580 took revenge, adding almost 8% to the average framerate of the RX 480 and overtaking the GeForce GTX 1060 6GB by 20%. Plus, Doom
Frame rate variation at 1920 x 1080 resolution is not a big problem for any card. However, Nvidia's attempt to cope with the Radeon RX 480 4GB leads to an unexpected result, with the GTX 1060 3GB falling behind the GTX 1050 Ti 4GB. You might argue that the three-gigabyte card wasn't designed for first-person shooters, but that's small consolation for those who buy it without knowing about these limitations.
AMD Radeon RX 580 20% faster than GeForce GTX 1060 6GB in Doom at 2560 x 1440 (RX 480 was 11% faster). Even the new Radeon RX 570 feels great at Ultra settings.
Meanwhile, the three-gigabyte GTX 1060, which should be quite competitive, continues to suffer from instability.
|
|||
|
| |||
Here's a basic, in-depth look at the AMD Radeon RX 580.
Object of study: 3D graphics accelerator (video card) AMD Radeon RX 580 8 GB 256-bit GDDR5
Developer Information: ATI Technologies (ATI trademark) was founded in 1985 in Canada as Array Technology Inc. In the same year it was renamed ATI Technologies. Headquarters in Markham (Toronto). Since 1987, the company has concentrated on producing graphics solutions for PCs. Since 2000, Radeon has become the main brand of ATI graphics solutions, under which GPUs are produced for both desktop PCs and laptops. In 2006, ATI Technologies was acquired by AMD, which formed the AMD Graphics Products Group (AMD GPG). Since 2010, AMD has abandoned the ATI brand, leaving only Radeon. AMD is headquartered in Sunnyvale, California, and AMD GPG remains headquartered in the former AMD office in Markham, Canada. We don't have our own production. The total number of AMD GPG employees (including regional offices) is about 2,000 people.
At a time when the PC components market as a whole is clearly not in its best shape, sales of gaming systems and related components are not only stable, but also managing to grow. It is not surprising that graphics chip manufacturers are trying to offer newer and newer video card models in order to improve their market positions and financial performance, even when there are no special new products and there is nowhere for them to come from. But you can always slightly overclock the GPU, install more video memory operating at a higher frequency, etc.
Moreover, the market for hardware intended for games is well supported from the software side - a large number of demanding game projects are coming out, they are switching to using new graphics APIs, with difficulty, but still entering the market of virtual reality systems that require more than double more GPU power. All this plays into the hands of manufacturers of graphics processors and video cards, who this year expect an increasingly widespread use of the DirectX 12 and Vulkan graphics APIs, and a many times greater number of releases of VR projects and sales of virtual reality helmets.
As usual, progress is most hampered by those who do not update their gaming systems and do not bring money to the manufacturers. AMD estimates that more than half a billion PC gamers still use graphics cards based on GPUs from two years ago or older, based on a March Steam Hardware Survey report. Only 10% of the total PC gamers upgraded their graphics subsystems last year, and about the same number use expensive video cards from previous generations, which are still quite relevant.
But as many as 80% of PC players have in their systems video cards of the Radeon R9 380X class and lower, which, although they provide more or less acceptable performance in modern games at Full HD resolution and not the highest quality settings, are still clearly inferior to new products and in terms of capabilities and performance. For example, the power of the Radeon R9 380 is not enough to provide 60 FPS at a resolution of 2560x1440 (and such monitors are now widespread) in modern gaming projects such as Battlefield 1, Doom and others:
In all the games presented in the diagram, the Radeon R9 380 video card is capable of providing a frame rate of only 41-45 FPS, which leads to an uncomfortable game, unsmooth video and problems with image tearing on the screen of a typical monitor. Accordingly, quality-demanding players have no choice but to lower the settings and rendering quality in order to get a stable 60 FPS and higher. Instead of upgrading your video card...
The same goes for virtual reality support, by the way. Only 18% of Steam users have video systems that meet the level of performance required for high-quality and comfortable VR (for this, a value of 6.0 or higher must be achieved in the SteamVR performance test). That is, 82% of users do not have enough GPU in their systems to try virtual reality in action.
In addition to performance, there is also support for new technologies, which owners of outdated video cards are deprived of. They won't be able to connect a modern HDR TV to their PC, for example. Or watch streaming video in 4K resolution with hardware decoding in the modern HEVC format. It is for such users who, for one reason or another, have not yet upgraded their video cards, AMD has released updated solutions of the Radeon RX 500 family, based on the already familiar Polaris chips.
Alas, we will have to wait a few more weeks for new top-end solutions based on the long-awaited Vega chips, using the second generation of very fast HBM memory, but for now let's look at the relaunch of the middle ones from the improved Polaris family. AMD believes that now is a great time to upgrade video cards, and has presented the public with new solutions that feature an attractive combination of performance, technology and price. Once again they are offering the market video cards that are less powerful than the best solutions of their competitors, but at attractive prices.
The new line is not much different from the existing one, except for the number 5 in the series name and the Radeon RX 550 model, which we will talk about a little later. What is the difference between the new products and the video cards that we already know? Is the Radeon RX 580 a complete copy of the Radeon RX 480? Yes and no. From a technical point of view, Polaris 20 is actually the same Polaris 10. But there are nuances. Firstly, the long-awaited 14 nm FinFET process has already significantly improved, which allowed AMD to significantly increase GPU clock speeds, while simultaneously reducing the prices of ready-made solutions, taking advantage of the reduction in the cost of chip production. According to rumors, even a different version of the technical process is used, but the details are not disclosed.
Secondly, some reference cards have changed a little, which we will discuss later, and their production partners have received expanded opportunities to create unique products. For example, if we take the Radeon RX 580, which we will consider today, the new model received not only increased GPU operating frequencies, but also changed characteristics for maximum power consumption, which should improve its capabilities when operating in extreme modes, including overclocking. But let's talk about everything in order...
Since the basis of the Radeon RX 580 model is the Polaris 20 graphics processor, which is essentially a renamed Polaris 10 chip and has a fourth-generation GCN architecture, which is similar to the architecture of previously released AMD solutions, before reading the theoretical part of the article it will be useful to familiarize yourself with our previous materials based on the company's video cards from previous generations:
- AMD Radeon RX 480: New mid-ranger, catching up with the top accelerators of the previous generation
- AMD Radeon R9 Fury X: AMD's new flagship with HBM support
- AMD Radeon R9 285: Tahiti received a 256-bit bus and turned into Tonga
- AMD Radeon R9 290X: Reach for Hawaii! You will achieve new heights of speed and functionality
- AMD Radeon HD 7970: New Single-GPU 3D Graphics Leader
Let's look at the detailed characteristics of the Radeon RX 580 video card, based on the full version of the Polaris 20 graphics processor.
| Graphics accelerator Radeon RX 580 | |
|---|---|
| Parameter | Meaning |
| Chip code name | Polaris 20 |
| Production technology | 14 nm FinFET |
| Number of transistors | 5.7 billion |
| Core area | 232 mm² |
| Architecture | Unified, with an array of common processors for stream processing of numerous types of data: vertices, pixels, etc. |
| DirectX hardware support | DirectX 12, supporting Feature Level 12_0 |
| Memory bus | 256-bit: 8 independent 32-bit memory controllers supporting GDDR5 memory |
| GPU frequency | 1257 (1340) MHz |
| Computing blocks | 36 GCN compute units, including 144 SIMD cores, consisting of a total of 2304 floating point ALUs (integer and floating point formats supported, with FP16, FP32 and FP64 precision) |
| Texturing blocks | 144 texture units, with support for trilinear and anisotropic filtering for all texture formats |
| Rasterization Units (ROPs) | 32 ROP blocks with support for anti-aliasing modes with programmable sampling of more than 16 samples per pixel, including with FP16 or FP32 frame buffer format. Peak performance up to 32 samples per clock, and in Z only mode - 128 samples per clock |
| Monitor support | Integrated support for up to six monitors connected via DVI, HDMI 2.0b and DisplayPort 1.3/1.4 Ready interfaces |
| Radeon RX 580 reference graphics card specifications | |
|---|---|
| Parameter | Meaning |
| Core frequency | 1257 (1340) MHz |
| Number of universal processors | 2304 |
| Number of texture blocks | 144 |
| Number of blending blocks | 32 |
| Effective memory frequency | 8000 (4×2000) MHz |
| Memory type | GDDR5 |
| Memory bus | 256-bit |
| Memory | 4/8 GB |
| Memory Bandwidth | 256 GB/s |
| Compute Performance (FP32) | up to 6.2 teraflops |
| Theoretical maximum fill rate | 43 gigapixel/s |
| Theoretical texture sampling rate | 193 gigatexels/s |
| Tire | PCI Express 3.0 |
| Connectors | One HDMI and three DisplayPorts |
| Energy consumption | up to 185 W |
| Additional food | One 8-pin connector |
| Number of slots occupied in the system case | 2 |
| Recommended price | 13 449 /15 299 |
The name of the video card announced today is consistent with AMD's current naming system. The model name differs from its predecessor RX 480 only in the changed number of the conditional generation - RX 5 80. The change, although understandable, is quite controversial, because in fact there is no new generation, it’s still the same good old Polaris. A name like Radeon RX 485 would be much more fair, but does it actually sound like it? However, AMD can be understood: given the absence of truly new GPUs for almost a year, they need to release something, and it is desirable that this one something there was a louder name.
It is clear that the senior model of the Radeon RX 500 family is taking the place of the previous RX 480, which is similar in positioning, in the company’s current line, because it is simply its overclocked version. The recommended prices for the Radeon RX 580 are quite attractive, but they are not much different from the prices of video cards almost a year ago. However, on the market, the new AMD video card will compete with the more expensive version of the GeForce GTX 1060 with 6 GB of video memory, and against its background the offer in terms of price and performance in its class can be called quite good.
Like their predecessors, Radeon RX 580 graphics cards will come in 4GB and 8GB GDDR5 memory versions, giving the company's partners maximum flexibility. The junior version will allow buyers to save money, since 4 GB is still sufficient for most games and is the optimal amount of video memory for a video card of this class. But the 8 GB option also has its advantages, which are already revealed in games when using high resolutions, and in the future this amount of video memory will only become more relevant. 8 GB of memory provides a good margin for the future, as VRAM requirements are constantly growing, and it will avoid unpleasant performance drops and FPS jerks.
For additional power, the reference board uses one 8-pin connector, unlike the 6-pin connector of the Radeon RX 480, and the typical power consumption for the new product is 185 W, as opposed to 150 W for its predecessor, and this is the right decision. It allows you to provide the GPU and other components with a large amount of power, which is especially important given the problems of the reference Radeon RX 480, which was forced to take a lot of power from the PCI Express slot, since the single 6-pin connector did not provide the amount of power necessary for the GPU.
It is clear that the power supply made it possible to improve both the frequency characteristics of the reference video card and its overclocking potential. By the way, the “closeness” of Polaris 20 and Polaris 10 played to the benefit of the new product here, since AMD’s partners immediately announced and released factory overclocked versions of the Radeon RX 580 video card presented today, differing from each other in both cooling systems, design and characteristics power subsystems. We are not even sure whether all partners noticed such a change in GPU in principle.
Architectural features
The Polaris 20 GPU belongs to the fourth generation of the Graphics Core Next architecture, the most advanced to this day. The Polaris architecture hasn't changed much from previous GPU generations, but improvements include: improved geometry handling, support for multiple views when rendering VR at different resolutions, an updated memory controller with improved data compression, modified instruction prefetching and improved buffering, scheduling and prioritization computing tasks in asynchronous mode, support for operations on data in FP16/Int16 format, etc.
The basic building block of the architecture is the Compute Unit (CU), from which all AMD GPUs are assembled. The CU compute unit has dedicated local data storage for exchanging data or extending the local register stack, as well as a read-write L1 cache and a full texture pipeline with fetch and filter units, it is divided into subsections, each of which works on its own stream of commands. Each of these blocks plans and distributes work independently. The GPU circuit is similar to Polaris 10 (click on the image to see a larger version of the illustration):
The full Polaris 20 GPU includes one Graphics Command Processor, four Asynchronous Compute Engines (ACE), two Hardware Schedulers (HWS), 36 Compute Units (CU), four geometry processors, 144 texture processors. TMU module (including four LSUs per TMU) and 32 ROPs. The memory subsystem of the AMD GPU in question includes eight 32-bit GDDR5 memory controllers, providing a total 256-bit memory bus, and a 2 MB L2 cache.
All improvements to the Polaris family of GPUs, including increased L2 cache and improved data caching, modified geometry engines, Primitive Discard Accelerator, new index cache for instanced geometry, instruction prefetching and caching, support operations on data in FP16 and Int16 formats, improvements to the task scheduler for asynchronous calculations, you can read in the review of the Radeon RX 480 video card.
We will focus on the differences between the Radeon RX 580 and the RX 480. There are not very many of them - in fact, these are only the GPU frequencies, which have increased from 1120/1266 MHz to 1257/1340 MHz for the base and turbo frequencies, respectively. This gives almost all peak performance numbers a 6%-12% boost, which isn't too bad for the same GPU, albeit an optimized one. Unfortunately, the same size GDDR5 memory still runs at the same 8GHz effective clock speed, which translates to 256GB/s of memory bus bandwidth, which in some cases can limit overall performance despite all the improvements in internal data compression. . There's not much more you can get out of GDDR5, and GDDR5X is probably too expensive for sub-$250 solutions.
There are no changes in the capabilities and support of new technologies in the Polaris 20 graphics chip, and the new Radeon RX 580 video card supports exactly the same as the Radeon RX 480 and other representatives of the previous series. This includes support for the latest DisplayPort and HDMI standards. The Radeon RX 400 family of graphics cards were among the first to support DisplayPort 1.3 HBR3 and DisplayPort 1.4-HDR, featuring increased bandwidth and support for HDR displays, and the Radeon RX 580 is no different. The same applies to support for hardware encoding and decoding of video data - Polaris has made some improvements in this area, which we have already written about. But the Radeon RX 580 also has something new - a new low-power mode.
Low Power Mode
Not only do all Radeon RX 500 series graphics cards offer superior power efficiency, but they also feature a low-power mode for video memory to help improve power efficiency when connecting to dual displays and when viewing multimedia content. The company's previous GPUs only support two states for video memory frequencies: maximum performance and idle. Connecting a second display to them automatically selects the performance mode, which leads to increased energy consumption.
The Radeon RX 500 family introduced a third, intermediate mode of video memory operation, and now the Radeon RX 580 consumes noticeably less power in such conditions compared to the Radeon RX 480. Thus, when connecting two 4K displays via DisplayPort, the RX 580 requires only 1. 5W more power compared to a single display-attached configuration, and the RX 480 uses almost 8W more in these conditions. We’re not even talking about the Radeon R9 380; it draws energy in big spoonfuls.
The same goes for power consumption when viewing video data - new solutions, including the Radeon RX 580, consume much less power than their predecessors in this mode. The graphics card we're reviewing today consumes only 14 W when streaming 1080p video from Youtube, compared to almost 39 W for the Radeon RX 480. Such achievements may not seem too significant, but when you consider that most of the time our PCs are in such modes , it becomes clear that it is in them that you can save the most energy.
Software technologies
AMD continues to improve the software component of its products. In particular, we can note the release of a new edition of the driver package in December 2016 - ReLive Edition. This edition has already been downloaded by users more than 10 million times, it brought new features, such as support for hardware encoding of 4K video in HEVC and H.264 formats for recording gameplay and streaming on popular services like Twitch. Naturally, hardware decoding of these formats is also supported.
Also in the new ReLive Edition, support for Radeon Chill technology has appeared, which allows, while maintaining a comfortable gaming experience, to slightly reduce image output delays in some cases and in all supported games to achieve greater energy efficiency, reducing energy consumption and the heating temperature of the GPU (and therefore the noise from the video card cooler ). In addition to a ton of gaming apps, Chill technology has recently appeared in two of the most popular games: League of Legends and DOTA 2.
The technology constantly monitors user input to determine the extent of user activity. And if it doesn't do anything, then Chill slightly reduces the frame rate so as not to consume unnecessary energy. But as soon as the user begins to act, Chill immediately increases FPS to a high level to maintain comfort while playing. And this is done instantly, the user does not feel anything. In addition, the technology eliminates too high frame rates, which stupidly consumes energy and does not provide any benefits. It is logical that the technology can bring the greatest benefit to games used in eSports.
By the way, if we talk about Radeon Chill technology in relation specifically to the Radeon RX 580, then the new product being considered today, when enabled, provides lower output latency in the Overwatch game at a resolution of 1920x1080 pixels and ultra-quality settings (6.1 ms versus 16.7 ms ) and power consumption (91 W vs. 126 W) while maintaining an acceptable frame rate of 91 FPS compared to 155 FPS in normal mode. Also, using Chill reduces the GPU temperature by several degrees.
But AMD measured such a reduction in energy consumption in several typical e-sports games. Depending on the game's demands on GPU power, enabling Chill technology can save up to 50%-70% energy! Of course, not every game is able to benefit from the considered technology for reducing energy consumption; it only works in the list of applications strictly defined by the company’s specialists, but there are already quite a few such projects, and the most important ones are there.
Preliminary performance evaluation and conclusions
So, we understand that the new model of the Radeon RX 580 video card is designed for the latest games at high resolutions and high quality settings, it supports the new DirectX 12 and Vulkan graphics APIs, is capable of providing maximum comfort in virtual reality applications, and also supports the company’s software technologies AMD, including FreeSync, ReLive and Chill.
But the most interesting question is the performance of the new product, because it is not so different from the Radeon RX 480 even in its theoretical characteristics. AMD compares the updated model not with the previous generation, but with older ones, since there is almost no difference and positions the Radeon RX 580 as a model for upgrading old solutions, so that players who decide to do so can get decent comfort in a resolution of 2560x1440 at high settings:
As you can see, while the Radeon R9 380X and GeForce GTX 970 video cards are often unable to show the 60 FPS required for comfort and smoothness in such conditions, the new Radeon RX 580 model can easily achieve performance of 69-81 FPS in the games presented in comparison. On average, according to AMD, the new product has 1.6 times more power than the Radeon R9 380X when it comes to gaming performance.
The same goes for virtual reality applications. The Radeon RX 580 not only provides greater performance in VR games compared to the Radeon R9 380, but also makes many of them playable in principle, because virtual reality no longer requires 60 frames per second to be comfortable, but 90!
If in the VR application SuperHot an outdated video card provides 92 FPS, which is enough for comfort (while the RX 580 achieved an impressive 141 FPS), then in Arizona Sunshine and Serious Sam VR the Radeon R9 380 video card is capable of providing only 72 and 78 FPS, respectively, which not enough for virtual reality. And this is not a joke; with a low frame rate in a VR helmet, not only slight discomfort is possible, but also tangible physical ailments in the form of dizziness associated with deception of the brain. So for those wanting to try out VR, a powerful GPU is a must, and the Radeon RX 580 is a good option for such an upgrade.
If we talk about product positioning and comparison with competitor solutions, the Radeon RX 580 model is positioned on the market against the GeForce GTX 1060 and looks good in this comparison - most often they are approximately equal, but sometimes the AMD video card has a slight advantage. Even more interesting is the comparison with various video cards of the past in a resolution of 2560x1440 pixels, which was carried out by company specialists:
It turns out that the Radeon RX 580 really looks like a good option for upgrading video systems at the level of the GeForce GTX 960 and Radeon R9 380 - performance in some cases can almost double. And even if we compare the new product with the GeForce GTX 970, which is still quite good in games, the Radeon RX 580 has a clear advantage in games that use modern graphics APIs: DirectX 12 and Vulkan. And given the presence of 8 GB of video memory in the new AMD model, its advantage will only get worse in the future.
The Radeon RX 580 is the most powerful member of the updated Polaris family, and is based on an optimized GPU manufactured using an improved 14 nm FinFET process technology. Compared to previous generations, architectural optimizations and technical processes have significantly improved the energy efficiency of the new product, but compared to the Radeon RX 480, there are, frankly, very few changes. Although they do exist: due to the increased GPU frequencies, you can expect a performance increase of 5%-10% in cases where it is not limited by video memory bandwidth.
The older Polaris remains AMD's best GPU at the moment, the list of functional changes and improvements is quite long, but... only in comparison with outdated GPUs of previous years. After all, Polaris 20 is the same Polaris 10, but with minor changes in the form of an improved power system (only for the reference card, partners also used enhanced and improved circuits on the Radeon RX 480), a new low-power mode and some software technologies applicable to all company decisions. Otherwise, it's the same Polaris that we've known for almost a year.
Thus, the main thing about the Radeon RX 580 video card presented today is that it is a good product in its class, quite suitable for upgrading the gaming PC of those users who, for some reason, have not yet done so. The new product is perfect for all modern games when installed at high or close to maximum quality settings and in relatively high resolutions, and it will also look at the proper level as part of virtual reality systems.
The combination of an attractive price and fairly high performance, slightly increased by increasing the clock speed of the graphics processor, makes the Radeon RX 580 one of the most successful video cards in the mid-price segment, attracting a large number of potential buyers. It is with these solutions that AMD usually wins back its market share. Although the new product has strong competitors, they are somewhat more expensive with similar performance indicators, and the issue of functionality is debatable; they all have their advantages.
In the following parts of the article, we will evaluate the performance of the new Radeon RX 580 video card in practice, comparing its speed with other accelerators from Nvidia and AMD. First, let's look at the data obtained in a small set of synthetic tests, and then move on to gaming tests.
| Average price (number of offers) in Moscow retail: | |
|---|---|
| Cards considered | Competitors |
| GTX 1060 6 GB - 16,500 rub. (as of 05/10/17) | |
| RX 580 8 GB - 19,000 rub. (as of 05/10/17) | GTX 1060 3 GB - 14,000 rub. (as of 05/10/17) |
| RX 580 8 GB - 19,000 rub. (as of 05/10/17) | RX 480 8 GB - 17,500 rub. (as of 05/10/17) |
| RX 580 8 GB - 19,000 rub. (as of 05/10/17) | GTX 970 4 GB - 17,500 rub. (as of 05/10/17) |
| Thermaltake DPS G 1050W power supply for the test bench provided by the company Thermaltake |
1. Review of the ASUS ROG Strix Radeon RX 580 8 GB video card (ROG-STRIX-RX580-T8G-GAMING)
technical characteristics of the video card and recommended costThe technical characteristics and cost of the ASUS ROG Strix Radeon RX 580 video card are shown in the table in comparison with the reference version of the AMD Radeon RX 580.
packaging and equipment
The video card was delivered to us for testing in a regular black cardboard box.
Therefore, we can only judge what the packaging of the ASUS ROG Strix Radeon RX 580 looks like based on photos from the official page of the video card.
The equipment, frankly speaking, is poor. In the box with the video card you will only find installation instructions, a DVD with drivers and two cable ties.
I don’t think that the elite ROG series video cards should be equipped like this in any way.
The video card is manufactured in China and comes with a three-year warranty. The cost of this model in Russia starts at 28,860 rubles, which is noticeably more expensive than the usual Radeon RX 580 options with 8 GB of memory.
PCB design and features
The design of the ASUS ROG Strix Radeon RX 580, like other video cards in this series, can be called aggressive. The massive plastic casing that covers the entire front side of the video card is not only curved in the horizontal plane, but also cut through with LED backlight inserts, and three fans with angular blades successfully fit their profile into the overall style of the video card.
The reverse side of the PCB is covered with a metal plate, but the cooler radiator and even individual elements on the PCB are visible from above and below.
The dimensions of the video card are 303 x 122 x 49 mm, and it weighs 1303 grams.
There are five video outputs on the rear panel: one DVI-D, two HDMI version 2.0b and two DisplayPort version 1.4.
The remaining area of this panel is perforated with slots for the outflow of air heated by the video card outside the system unit case.
Additional power for the ASUS ROG Strix Radeon RX 580 is provided by one eight-pin connector, soldered in the usual place.
The power consumption of the video card stated in the specifications is standard for 185 watts, and the recommended power supply power for a system with one such video card should be at least 500 watts.
We show the design of the ASUS ROG Strix Radeon RX 580 in the following diagram from the ASUS website.
Despite its massiveness, the cooling system is secured to the printed circuit board with only four screws around the perimeter of the GPU.
To completely “undress” the video card, you will have to unscrew the heat distribution plate on the front side and the protective plate on the back. This is what the ASUS ROG Strix Radeon RX 580 looks like without all its “armor”.
The graphics card's printed circuit board is made using fully automated Auto-Extreme technology, which allows for higher quality and increased component reliability.
The GPU power system, built using Super Alloy Power II technology, uses six phases, and two more are allocated to video memory and power circuits.
The DIGI+ VRM ASP1300 controller is responsible for managing the power of the GPU.
Two BIOS chips are soldered next to the backlight connector on the back of the video card.
At the end of the video card opposite the outputs there are connectors for connecting fans and lighting, as well as two connectors for connecting external fans, the rotation speed of which will be controlled by the video card itself (FanConnect II technology).
Near the additional power connector there are points for measuring voltages.
The ITE8295FN controller on the PCB board is responsible for adjusting the backlight of the video card.
The ASUS ROG Strix Radeon RX 580 has a backlit fan shroud on the front side and a plate on the back of the board.
Six backlight modes and any color to choose from are available for customization.
The 14nm Polaris 20 XTX GPU has a die area of 232mm2, which is average by today's standards. The GPU of our video card was released in the 9th week of 2017 (late February - early March).
In 3D mode, the ASUS ROG Strix Radeon RX 580 processor can operate at frequencies of 1411 or 1431 MHz (OC mode), which makes this model one of the fastest among the factory Radeon RX 580 (+6.8% to the reference value). When the video card switches to 2D mode, the GPU frequency is reduced to 300 MHz with a simultaneous reduction in voltage from 1.112 V to 0.750 V.
Eight gigabytes of GDDR5 memory are equipped with Samsung chips labeled K4G80325FB-HC25.
The theoretical effective frequency of such chips is 8000 MHz, at which the ASUS ROG Strix Radeon RX 580 memory operates, which, with a memory bus width of 256 bits, provides a throughput of 256 GB/s. In 2D mode, the memory frequency is reduced to 1200 MHz.
At the end of this subsection of the article, we present the characteristics of the ASUS ROG Strix Radeon RX 580 from the GPU-Z utility.
cooling system - efficiency and noise level
ASUS ROG Strix Radeon RX 580 is equipped with a new cooling system, which for some reason has not received any name. But we can say unequivocally that this is not some variation of the DirectCU cooler with a direct contact base, which received a lot of criticism. Now the cooler has a classic base and a larger radiator.
The thickness of the radiator fins has been increased by 40%, and now the video card occupies two and a half slots instead of two, with such an area it can handle any thermal loads.
The base of the radiator is pierced by six nickel-plated copper heat pipes with a diameter of 6 mm. Three of them distribute the heat flow along the ribs of the main section of the radiator, and five - over the additional section.
However, all six heat pipes are also in contact with the main section at its base. In other words, the radiator developers tried to make maximum use of heat pipes in its design.
In addition to increasing the area, the key innovation in the radiator is MaxContact technology, which consists of polishing the base to a mirror finish.
According to the developers, such a base should ensure faster heat transfer between the GPU crystal and the heatsink. There is only one question left - this is a long-known fact, especially in the case of open processor chips, then why couldn’t this have been done earlier?
The finishing quality of the radiator base is truly amazing.
The elements of the power supply circuit of the graphics processor and memory, as well as the video memory chips themselves, are cooled by a metal plate approximately 1.5 mm thick, which is in contact with them through thermal pads. The plate on the back side of the PCB performs only a protective and decorative function and does not participate in cooling the video card.
To cool the radiator, three fans with a diameter of 87 mm are installed on it. It uses proprietary impellers with an aggressive blade shape and an end fin, due to which increased static pressure is achieved and air flow increases. In addition, the fans meet the IP5X standard (partial dust protection). Their rotation speed is controlled by pulse-width modulation in the range from 0 to 3510 rpm (according to monitoring data). That is, when the GPU temperature is below 55 degrees Celsius, the fans stop. In my opinion, this is not a very good solution, since no one will hear these fans in the system unit case already at 700-800 rpm, but the video card will operate at temperatures 15 degrees Celsius lower and warm up under load much longer.
To check the temperature conditions of the video card as a load, we used nineteen cycles of the Fire Strike Extreme stress test from the 3DMark graphics package.
To monitor temperatures and all other parameters, MSI Afterburner version 4.4.0 Beta 15 and GPU-Z utility version 2.4.0 were used. Tests were carried out in closed system case, the configuration of which you can see in the next section of the article, at room temperature around 22,0 degrees Celcius.
First, let's check the efficiency of the ASUS ROG Strix Radeon RX 580 cooling system in automatic fan speed control mode.
Automatic mode (0-1620 rpm)
Well, everything is just wonderful here. The larger radiator area and perfectly polished base definitely benefited the new ASUS cooling system. Now it cools the video card so confidently that it does not allow the GPU temperature to rise above 65 degrees Celsius, and at the same time only 1620 rpm fan speed is enough. In our opinion, these are excellent indicators for such a hot video card with a good factory core overclock.
At the maximum speed of the three ASUS cooler fans, the GPU temperature decreased by 15 degrees Celsius compared to the automatic adjustment mode.
Maximum speed (~3510 rpm)
As you can see, the new ASUS cooler has excellent potential, and such temperatures are more typical for liquid cooling systems than air coolers. Now about the noise level.
The noise level of video card cooling systems was measured with an electronic sound level meter OKTAVA-110A after one in the morning in a completely closed room of about 20 m² with double-glazed windows. The noise level was measured outside the system case, when the source of noise in the room was solely the video card cooling system and its fans. The sound level meter, fixed on a tripod, was always located strictly at one point at a distance of exactly 150 mm from the fan rotor/cooler turbine. The motherboard, into which a video card was inserted with a cooling system installed on it, was placed at the very corner of the table on a polyurethane foam backing. The lower measurement limit of the sound level meter is 22 dBA, and the subjectively comfortable (please do not confuse with low) noise level when measured from such a distance is around 36 dBA. The fan rotation speed was varied over the entire range of their operation using a special precision controller by changing the supply voltage in 0.5 V steps.
In the graph with data on the noise level of the ASUS ROG Strix Radeon RX 580 cooler, we included the results of measurements of the noise level of cooling systems of previously tested video cards Inno3D GeForce GTX 1070 HerculeZ Twin X2 and Gigabyte GeForce GTX 1080 G1 Gaming. Let us add that vertical dotted lines of the corresponding color mark the upper limits of the speed of fans of cooling systems during their automatic PWM adjustment.
Interestingly, the noise level curve of the ASUS ROG Strix Radeon RX 580 video card cooler practically copies the noise level curve of the Inno3D GeForce GTX 1070 HerculeZ Twin X2 cooling system, but due to the lower fan speed in automatic adjustment mode, ASUS is quieter, although not quiet and not even comfortable. However, with such a high efficiency of the new cooler from ASUS ROG Strix Radeon RX 580, it is quite possible to achieve quiet operation at 1200-1250 rpm, albeit with a non-critical increase in the temperature of the GPU to 70-72 degrees Celsius.
overclocking potential
AMD Radeon RX 580 video cards do not indulge users with high overclocking potential, and even more so the factory overclocked versions. Unfortunately, the ASUS ROG Strix Radeon RX 580 was not a pleasant exception to these statistical observations, since its graphics processor only “overclocked” by 5-10 MHz without increasing the voltage. By increasing the core voltage by 48 mV, we were able to achieve stability at a frequency of 1465 MHz, which is also a very modest figure. But the memory did not disappoint, accelerating by 1000 effective megahertz or +12.5%.
As a result, the frequencies of the overclocked video card were 1465/9000 MHz.
During overclocking, we did not trust the automatic fan speed adjustment, fixing them at 62% power or 2500 rpm.
62% power (~2500 rpm)
Therefore, the temperatures of the overclocked ASUS ROG Strix Radeon RX 580 were predictably lower than in the nominal operating mode.
2. Test configuration, tools and testing methodology
Testing of video cards was carried out on a system with the following hardware configuration:
motherboard: ASRock Fatal1ty X299 Gaming K6 (Intel X299 Express, LGA2066, BIOS L1.30A beta dated 08/15/2017);
CPU: Intel Core i9-7900X (14 nm, Skylake-X, U0, 3.3-4.5 GHz, 1.1 V, 10 x 1 MB L2, 13.75 MB L3);
CPU cooling system: Noctua NH-D15 (2 NF-A15, 800~1500 rpm);
thermal interface: ARCTIC MX-4 (8.5 W/(m*K);
RAM: DDR4 4 x 4 GB Corsair Vengeance LPX 2800 MHz (CMK16GX4M4A2800C16) (XMP 2800 MHz/16-18-18-36_2T/1.2 V or 3000 MHz/16-18-18-36_2T/1.35 V) ;
video cards:
AMD Radeon RX Vega 64 8 GB/2048 bit 1630/1890 MHz;
Inno3D GeForce GTX 1070 HerculeZ Twin X2 8 GB/256 bit 1506-1683(1860)/8008 MHz;
ASUS ROG Strix Radeon RX 580 8 GB/256 bit 1411/8000 MHz;
NVIDIA GeForce GTX 1060 Founders Edition 6 GB/192 bit 1506-1709(1886)/8008 MHz;
disk for system and games: Intel SSD 730 480 GB (SATA III, BIOS vL2010400);
benchmark drive: Western Digital VelociRaptor 300 GB (SATA II, 10000 rpm, 16 MB, NCQ);
archive disk: Samsung Ecogreen F4 HD204UI 2 TB (SATA II, 5400 rpm, 32 MB, NCQ);
sound card: Auzen X-Fi HomeTheater HD;
case: Thermaltake Core X71 (five be quiet! Silent Wings 2 (BL063) at 900 rpm);
control and monitoring panel: Zalman ZM-MFC3;
Power supply: Corsair AX1500i Digital ATX (1500 W, 80 Plus Titanium), 140 mm fan;
monitor: 27-inch Samsung S27A850D (DisplayPort, 2560 x 1440, 60 Hz).
Today we will evaluate the performance of the ASUS ROG Strix Radeon RX 580 in comparison with the reference version of the AMD Radeon RX Vega 64 and the original Inno3D GeForce GTX 1070 HerculeZ Twin X2.
In addition to these video cards, the testing included NVIDIA GeForce GTX 1060 Founders Edition with 6 GB of memory.
Note that the power and temperature limits on video cards were increased to the maximum possible, maximum performance was prioritized in the GeForce drivers, and Turbo mode was activated in the Crimson driver for AMD Radeon RX Vega 64 and the HBCC (High Bandwidth Cache Controller) function was enabled. Since manual fan speed increase on the AMD Radeon RX Vega 64 was not activated, the graphics card's GPU operated at frequencies of 1460-1532 MHz.
To reduce the dependence of video card performance on the platform speed, the 14-nm ten-core processor with a multiplier of 44, a reference frequency of 100 MHz and the Load-Line Calibration function activated to the second level was overclocked to 4.4 GHz across all cores simultaneously when the voltage in the motherboard BIOS is increased to 1.11 V.
At the same time, 16 gigabytes of RAM operated in four-channel mode at a frequency 3.0 GHz with timings 16-16-16-28 CR1 at a voltage of 1.35 V.
Testing, which began on September 23, 2017, was carried out under the Microsoft Windows 10 Pro operating system (1703 15063.608) with all updates as of the specified date and with the installation of the following drivers:
motherboard chipset Intel Chipset Drivers – Intel Management Engine Interface (MEI) – href=http://station-drivers.com/index.php?option=com_remository&Itemid=353&func=fileinfo&id=3057&lang=fr" data-desc="11.7.0.1037 WHQL from 09/14/2017" >11.7.0.1037 WHQL from 09/14/2017;
Video card drivers for NVIDIA GPUs – GeForce 385.69 WHQL from 09/21/2017;
Video card drivers on AMD GPU – Crimson ReLive 17.9.2 beta from 09/21/2017;
In today's testing, we only used a resolution of 2560 x 1440 pixels. For the tests, two graphics quality modes were used: Quality + AF16x - texture quality in the drivers by default with 16x level anisotropic filtering enabled, and Quality + AF16x + MSAA 4x/8x with 16x level anisotropic filtering enabled and 4x or 8x full-screen anti-aliasing. In some games, due to the specifics of their game engines, other anti-aliasing algorithms were used, which will be indicated further in the methodology and in the diagrams. Anisotropic filtering and full-screen anti-aliasing were enabled directly in the game settings. If these settings were not available in games, then the parameters were changed in the control panels of the GeForce or Crimson drivers. Vertical synchronization (V-Sync) was also forcibly disabled there. Apart from the above, no additional changes were made to the driver settings.
In addition to the traditional update of all games in the test set, we excluded from it outdated or no longer relevant DiRT Rally, Gears of War 4 and Battlefield 1. Instead, Tom Clancy's Ghost Recon Wildlands and the new F1 2017 were included. As a result, the video cards were tested in two graphics tests and fourteen games, updated to the latest versions as of the start date of preparation of the material.Now the list of test applications is as follows (games and further test results in them are arranged in the order of their official release):
3DMark(DirectX 9/11/12) – version 2.3.3732, testing in the Fire Strike, Fire Strike Extreme, Fire Strike Ultra and Time Spy scenes (the diagram shows the graphical score);
Unigine Superposition(DirectX 11) – version 1.0, tested in 1080P High, 1080P Extreme and 4K Optimized settings;
Crysis 3(DirectX 11) – version 1.3.0.0, all graphics quality settings to maximum, blur level to medium, glare on, modes with FXAA and MSAA 4x, double sequential pass of a scripted scene from the beginning of the Swamp mission lasting 105 seconds;
Metro: Last Light(DirectX 11) – version 1.0.0.15, used the test built into the game, graphics quality settings and tessellation at the Very High level, Advanced PhysX technology activated, tests with SSAA and without anti-aliasing, double sequential run of the D6 scene;
Grand Theft Auto V(DirectX 11) – build 1180.1, quality settings at Very High, ignore suggested restrictions enabled, V-Sync disabled, FXAA activated, NVIDIA TXAA disabled, MSAA for reflections disabled, NVIDIA soft shadows;
Batman: Arkham Knight(DirectX 11) – version 1.6.2.0, quality settings at High, Texture Resolutioin normal, Anti-Aliasing on, V-Sync disabled, tests in two modes – with and without the activation of the last two NVIDIA GameWorks options, double sequential run of the built-in into the test game;
Tom Clancy's Rainbow Six: Siege(DirectX 11) – version 2.2.2, texture quality settings at Very High, Texture Filtering – Anisotropic 16X and other maximum quality settings, tests with MSAA 4x and without anti-aliasing, double sequential run of the test built into the game.
Rise of the Tomb Raider(DirectX 12) – version 1.0 build 770.1_64, all parameters for the Very High level, Dynamic Foliage – High, Ambient Occlusion – HBAO+, tessellation and other quality improvement techniques are activated, two test cycles of the built-in benchmark (Geothermal Valley scene) without anti-aliasing and with SSAA 4.0 activation;
Far Cry Primal(DirectX 11) – version 1.3.3, maximum quality level, high-resolution textures, volumetric fog and shadows to maximum, built-in performance test without anti-aliasing and with SMAA Ultra activated;
Tom Clancy's The Division(DirectX 11) – version 1.7.1, maximum quality level, all image enhancement parameters are activated, Temporal AA – Supersampling, testing modes without anti-aliasing and with SMAA 1X Ultra activated, built-in performance test, but FRAPS results are recorded;
Hitman(DirectX 12) – version 1.12.1, built-in test with graphics quality settings at “Ultra”, SSAO enabled, shadow quality “Ultra”, memory protection disabled;
Deus Ex: Mankind Divided(DirectX 12) – version 1.19 build 801.0, all quality settings are manually set to the maximum level, tessellation and depth of field are activated, at least two consecutive runs of the benchmark built into the game;
For Honor(DirectX 11) – version 32.175, maximum graphics quality settings, volumetric lighting – MHBAO, dynamic reflections and blur effect enabled, oversampling anti-aliasing disabled, tests without anti-aliasing and with TAA, double sequential run of the benchmark built into the game;
Tom Clancy's Ghost Recon Wildlands(DirectX 12) – version 1.6.0, graphics quality settings to maximum or Ultra level, all options activated, tests without anti-aliasing and with SMAA+FXAA, double sequential run of the benchmark built into the game;
Warhammer 40,000: Dawn of War III(DirectX 11) – version 4.320.2829.17945, all graphics quality settings at the maximum level, double sequential run of the benchmark built into the game;
F1 2017(DirectX 11) – version 1.6, the test built into the game was used on the Marina Bay highway in Singapore in a downpour, graphics quality settings were set to the maximum level for all points, SSRT shadows were activated, tests with TAA and without anti-aliasing.
Let's add that if games implement the ability to record the minimum number of frames per second, then it is also shown in the diagrams. Each test was carried out twice; the best of the two values obtained was taken as the final result, but only if the difference between them did not exceed 1%. If the deviations of the benchmark runs exceeded 1%, then the testing was repeated at least once more to obtain a reliable result.
3. Video card performance test results
3DMark
Unigine Superposition
Crysis 3
Metro: Last Light
Grand Theft Auto V
Batman: Arkham Knight
We will analyze individual results using summary diagrams.
4. Summary charts and analysis of results
In the first summary chart, we propose to compare the performance of the ASUS ROG Strix Radeon RX 580 and the new flagship AMD Radeon RX Vega 64 video card. Or rather, evaluate the lag of the heroine of today’s article from the fastest single-processor video card on an AMD graphics processor.
Only the results in the game Metro: Last Light fall out of the general row, but this is explained by the activation of the Advanced PhysX function, which all AMD video cards cannot cope with. As for other games, on average for all games without anti-aliasing, the Radeon RX 580 is inferior to the Radeon RX Vega 64 by about 33.3%, and almost the same (32.7%) with it activated. It would seem that the difference is quite significant, if not for one “but”. The fact is that the cost of these video card models today is such that for the price of one reference Radeon RX Vega 64 you can buy two original Radeon RX 580 8 GB and still have some left over on the HDD. By installing a couple of these video cards in CrossFireX mode, in most games you can get 30-40% higher performance than a single Radeon RX Vega 64. That's something like that.
Now let's evaluate how the ASUS ROG Strix Radeon RX 580 8 GB compares to the Inno3D GeForce GTX 1070 HerculeZ Twin X2 8 GB.
And again there is a lag, although in smaller percentages than in the case of the flagship AMD video card: 21.6% without anti-aliasing and 23.1% with it activated. Moreover, we have GeForce 1070 at reference frequencies, and Radeon RX 580 with factory overclocking. If we turn again to the cost of video cards, prices for eight-gigabyte Radeon RX 580 start at 22 thousand rubles, and GeForce GTX 1070 cost from 31 thousand. So, in this case, in our opinion, everything is quite natural.
Now let's look at the ASUS ROG Strix Radeon RX 580 8 GB in comparison with the reference NVIDIA GeForce GTX 1060 Founders Edition 6 GB.
Here the video card based on the AMD graphics processor is already a victory. Not counting modes using anti-aliasing in the games Crysis 3 and GTA V, as well as Tom Clancy's Ghost Recon Wildlands and Far Cry Primal, it is confidently ahead of the reference GeForce GTX 1060. On average, for all games, the advantage of the ASUS video card is 10.3% without use anti-aliasing and 8.2% when using various image quality improvement techniques.However, it is worth remembering that the GeForce GTX 1060 wins in terms of cost in this comparison, and such video cards overclock better than the Radeon RX 580 even without increasing the voltage on the GPU.
In the fourth diagram we will present a comparison that goes beyond the topic of today's article, namely, we will pit the AMD Radeon RX Vega 64 and GeForce GTX 1070 against each other. Since the AMD video card came out much later than NVIDIA, we took the GeForce GTX indicators as the starting point in the diagram 1070, and the results of Radeon RX Vega 64 were reflected in the form of deviations from them.
The GeForce GTX 1070 won in Metro: Last Light and was on par with the Radeon RX Vega 64 in Crysis 3 and GTA V using anti-aliasing. In other games, the AMD video card is faster, and on average its advantage is 18.7% without anti-aliasing and 16.1% with it activated. It's safe to say that the new GeForce GTX 1070 Ti should correct this situation.
5. Power consumption
The energy consumption level was measured using a Corsair AX1500i power supply via the Corsair Link interface and the program of the same name version 4.8.3.8. The energy consumption of the entire system as a whole was measured, excluding the monitor. The measurement was carried out in 2D mode during normal work in Microsoft Word or Internet surfing, as well as in 3D mode. In the latter case, the load was created using four consecutive cycles of the introductory scene of the Swamp level from the game Crysis 3 in a resolution of 2560 x 1440 pixels at maximum graphics quality settings using MSAA 4X. CPU power saving technologies are disabled in the motherboard BIOS.Let's evaluate the level of power consumption of systems with video cards tested today using the results in the diagram.
In terms of power consumption, video cards on AMD GPUs are inferior to their competitors, and the Radeon RX Vega 64 is quite a lot. The maximum power consumption level of the configuration with ASUS ROG Strix Radeon RX 580 8 GB is 32 watts higher than the system with Inno3D GeForce GTX 1070 HerculeZ Twin X2 and almost 100 watts higher than the system with NVIDIA GeForce GTX 1060 Founders Edition. It should be noted that in 2D modes, video cards with AMD GPUs turned out to be more economical than video cards with NVIDIA GPUs.
Conclusion
The ASUS ROG Strix Radeon RX 580 video card overall left a pleasant impression. The most striking of them is the new cooling system with a larger radiator and a perfectly polished base. Thanks to this, the video card operates at a relatively low temperature and leaves room for manually adjusting the fan speed in order to reduce noise levels. We have no comments about the hardware of the video card's printed circuit board. A powerful GPU power system with durable components and an automatic build mode should ensure that the ASUS ROG Strix Radeon RX 580 runs smoothly under the heaviest loads. The backlight of the video card is a pleasant and effective bonus to all of the above. At the same time, we cannot help but note the disadvantages of this product, including large dimensions and weight, relatively high cost and modest overclocking potential.
We thank AMD and personally Ivan Maznev
for the video card provided for testing.
The RX 580 may be a rebrand, but it's no slouch.
So let's start with the important points. This is not Vega. No heady heights of next-generation 14 nm architecture for you. This is not a new flagship that has finally joined us after a long absence from AMD; we have not seen high-end competition in the graphics card market for almost two years. No, in general, this is a renamed card. Or at least that's what the critics will tell you.
- pros: 1080p performance | Impressive price | Sapphire coolers;
- Minuses: GTX 1060 3GB is more valuable | Not Vega;
What do we mean by "renamed"? Well, to put it simply, you take the established architecture of Polaris and the RX480. You go back to the drawing board and optimize GPUs on silicon. You will improve the procedural part. After all, you take the best of the best architecture based on the same manufacturing process, bump up the clock speeds, and resell the card, usually as a lower-tier model than before.
Traditionally, we see such modernization when a new architecture is released, aimed at the middle and high-end consumers. And, as a rule, repeated re-releases are typical only for the budget range of video cards. The top two tiers of video cards are based on a new architecture, then the highest-class generation comes out, and against the backdrop of all this, a reissue comes out in the budget segment. And all this makes the RX 580 strange, since there is no serious flagship from AMD, there are no new designations, architecture, the video card is simply overclocked and returned to the market at the same price as its older brother.
CharacteristicsRadeonRX 580
Here's the Sapphire Radeon RX 580 Nitro+ configuration from our review:
- Universal processors: 2304;
- Texture blocks: 144;
- Rasterization units: 32;
- Clock frequency: 1340 MHz;
- Boost frequency: 1411 MHz;
- Video memory: 8 GB GDDR5;
- Memory frequency: 8 GHz;
- Memory bus: 256-bit;
- Nutrition: 1x8, 1x6;
- Interfaces: DVI-I, HDMI, DisplayPort;
Don't believe me? Well, let's compare them with each other. The RX 580 8GB (and RX 480 8GB) each offer 2,304 general-purpose processors (stream processors), 32 ROPs, 144 texture units, 8 Gbits of memory, a 256-bit memory bus, and one GCN 4.0 architecture.
The only difference that accompanies the video cards is the base clock speed and TDP, the RX 580 requires 185 W, compared to the 480 and 150 W, and the frequency boost is no more than 137 MHz, at least on the native board. Of course, this data can be converted by hardware partners like Sapphire, ASUS, MSI and others, but you won't see much difference from the original numbers.
PerformanceRadeonRX 580
You might conclude that the RX 580 is dead on arrival, not at all. This is a video card that is not intended to shake up the established ecosystem, just to balance it. As we've already seen with the big 14nm and 28nm releases, 100MHz significantly improves performance. Even a slight increase in clock speed has an impact on performance across the board.
Side by side, the Sapphire RX 580 8GB Nitro+ easily beats the latest RX 480 cards when it comes to 1080p and 1440p performance.
On average, the card gains 12% performance across all gaming benchmarks, with a staggering 1,500 point gain on Firestrike benchmarks relative to the RX 480. To provide perspective, the gain is equivalent to going from a GTX 1070 to a .
Should we assume that proud RX 480 owners will change ship for a new graphics card? Never. In fact, if you're itching for extra framerate, it's not worth spending the money, you could overclock the native 480 to 580 speeds and get similar benefits.
The AMD Radeon RX 580 is ultimately for those stuck with older Nvidia GPUs, 700 series, or even 300 series AMD graphics cards. This is the upgrade route for those looking to upgrade an old graphics card. If you want to play the latest games at 1080p, the new RX 580 is likely one of the best value propositions available today.
Design and Cooling
But what about our video card? Well, the Sapphire Radeon RX 580 Nitro+ is a real work of art.
The dual-fan cooling system has always been good when it comes to reducing noise levels as well as overall cooling potential, which is evident once you look at the 580's increased base clocks.
The inclusion of a black and white back panel will highlight any build. Add in a simple blue LED logo and Sapphire becomes a very attractive solution.
However, this is far from a bulky graphics card. In fact, we mean that the map is a good size. It still takes up two slots, but the overall dimensions of this configuration seem ridiculous compared to its competitors. Moreover, we are increasingly seeing compact configurations from Gigabyte and the like.
What's impressive is the size relative to the increase in base frequency. The 1340 MHz output is quite a significant offering, and you'll easily notice a 10% increase in gaming performance over the RX 480.
Summing up
Ultimately, the RX 580 is the graphics card for those who want to buy a graphics card that can run AAA games at 1080p resolution.
And while the RX 580 clock speed increase is not too great since the last generation 480s, maintaining the price tag of the previous model, raising the price-quality ratio.
And while the graphics card doesn't offer the best price/performance ratio on the market, for the performance it offers, the price is very good.
For those looking for better performance? We can only hope that Vega is closer than we thought. Otherwise, AMD's re-release strategy doesn't make much sense since the manufacturer is only adding a little bit of performance from the previous card.
AMD Radeon RX 580 Review
Anton Zim
21.04.2017 This is undoubtedly not the graphics card we were expecting from AMD right now. The RX 580 isn't exactly a Titin killer and isn't going to change the graphics card ecosystem in any significant way, but it's a serious contender for the Nvidia GTX 1060's mid-range crown.
8 Overall Score
Verdict:
This is undoubtedly not the graphics card we were expecting from AMD right now. The RX 580 isn't exactly a Titin killer and isn't going to change the graphics card ecosystem in any significant way, but it's a serious contender for the Nvidia GTX 1060's mid-range crown.
Loading...
