Radeon r7 240 2gb games. AMD Radeon R7 and R9 - an updated line of graphics cards

Today we will talk about a series AMD graphics cards Radeon R7 200 series. We will talk about four representatives: 260, 250 and 240 series. Despite the fact that there are several individual options, the difference between them is not so great. We will definitely point out and consider all the most noticeable differences so that you do not have to think again about what to buy.

Price

Let's start right away with the price category of these cards. Despite the fact that they are all high-performance graphics processors and allow you to easily process large data streams, they are in the so-called economy sector. On average, you will have to pay up to 10,000 rubles for such a card, depending on the specific model and store, which is good news.

Although modern games are created exclusively for the latest video cards, and ultra-settings work only on video processors from 50,000 rubles, this does not mean that cheap cards cannot replace them. The fact is that the AMD Radeon R7 200 series reviews from users are extremely positive, which means they are in high demand.

Characteristics

We'll start with the most common 200 series. All of them are manufactured according to the technical process that determines the crystal size of 28 nm. The value is not too high compared to other video cards, it could be better. Otherwise, this would lead to overheating of the card, but two built-in cooling coolers negate this design flaw in the AMD Radeon R7 200 series. The photos in the article clearly demonstrate this. The only drawback is that you will have to clean the cooling system more often.

AMD Radeon R7 200 series, the characteristics of which we are considering, are connected via the PCI-E x16 version 3.0 interface. Quite standard, but at the same time much better than AGP. When buying, just pay attention to this nuance.

Monitor

As for displaying information on the screen, then the AMD Radeon R7 200 series driver is hot. If in the 240 series the video card supports only two monitors, then in all subsequent ones there are several possibilities.

• If you use HDMI as well, you can connect up to 3 screens.
• With the DisplayPort connector, their number increases to four.
• When using an MST hub, you can connect as many as 6 monitors.

At the same time, these video cards support a resolution of 4096x2160. These video cards are suitable for both conventional and widescreen displays. Thus, video cards are able to support multi-monitor systems, and software, supplied with them, will help the user with the implementation of the plan.

performance characteristics

Let's now talk about the technical component of the AMD Radeon R7 200 series. The characteristics of these video cards are such that they are able to support the most modern and demanding games without much effort. The exception is AMD Radeon But it was released a long time ago.

For example, the frequency of the GPU, which largely determines the performance of a video card, fluctuates around 1 GHz and may vary depending on the manufacturer. At the same time, for the 240 model, this value is approximately 800 MHz.

The memory of the video card has GDDR5 and DDR3 formats. But at the same time, if the old models use only outdated memory, then the new ones (260 series) contain exclusively modern technology.

It can also vary significantly within the same model. It also depends entirely on the manufacturer. However, despite your best efforts, you will not be able to find an AMD Radeon R7 200 series graphics card that has more than 2 gigabytes of memory. But this is not necessary if you are interested in modern games and their requirements.

Based on the memory format, the bandwidth of the AMD Radeon R7 200 series also follows. The characteristics that we get at the output demonstrate the working abilities of these video cards in the best possible way:

• 250 and 240 series have a bandwidth of up to 72 GB / s;
• 260, 260x and 265 provide data transfer rates of 96/104/180 GB / s, respectively.

As you can see, with many other parameters being equal, only the newest models win. However, with an average price of 7700 rubles, buying this video card to upgrade your PC will not be a big problem.

Computing process

Having considered the general data available to the buyer on the price tag in the store, we move on to more accurate figures. The characteristics of the AMD Radeon R7 200 series will allow us to say exactly which of these video cards is better than the rest.

Let's start with the number of universal processors. With their help, the color and shape of objects displayed on the screen are calculated. As you understand, the performance of the card as a whole greatly depends on this parameter. And here we can safely say that the 240 model is far behind its counterparts in the series.

• AMD Radeon R7 240 has only 5 compute units, which is 320 processors.
• AMD Radeon R7 250 contains 8 blocks, and 250x already contains 14.
• The AMD Radeon R7 260 has 12 blocks, the 260x has 14, but the 265 model contains as many as 16 computing devices. Explanations are not even required here.

Additionally

Of course, the technical characteristics of the AMD Radeon R7 200 series are very important. What technologies a modern video card can support determines its functionality and usefulness for the user.

Video cards of the reviewed series feel confident when working with Direct X 12.0 and Open CL 1.2. They also work great with Open GL 4.3. And the CrossFire technology has long ceased to amaze and is also perfectly supported by these cards. What then are the differences?

1. The first stumbling block is AMD TrueAudio. This technology was created to maintain a higher quality sound. That's just not all games use it, just like not all equipment is able to support it. Of all the presented cards, this technology is supported only by 260 and 260x.
2. Also, the VCE decoder responsible for HD video can become a problem. It is supported only in cards from 250x and above.

We continue the series of articles about video cards of the lower price segment. In the previous material, the features of the GeForce GT 730 with different types of memory were considered. Now let's talk about AMD graphics solutions.

The main graphics processor in AMD's low-end segment is the Oland GPU, made using the 28nm process technology. It has 384 stream processors, 32 texture units, 8 ROPs and a 128-bit memory bus. The processor is the main one for the Radeon R7 250, and a stripped-down version is used in the Radeon R7 240. The older video card has already been reviewed by us in one of the old materials. In this article, we will get acquainted with a simpler version that uses DDR3 memory. At the same time, we will review the younger video card Radeon R7 240, the potential of which is limited to 320 stream processors and a lower core frequency.

Video adapter	Radeon R7 250 GDDR5	Radeon R7 250 DDR3	Radeon R7 240 GDDR5	Radeon R7 240 DDR3
Nucleus	Oland XT	Oland XT	Oland Pro	Oland Pro
	1040	1040	1040	1040
Process technology, nm	28	28	28	28
Core area, sq. mm	90	90	90	90
	384	384	320	320
Number of texture blocks	24	24	20	20
Number of render units	8	8	8	8
Core frequency, MHz	1000-1050	1000-1050	730-780	730-780
Memory bus, bit	128	128	128	128
Memory type	GDDR5	DDR3	GDDR5	DDR3
Memory frequency, MHz	4600	Before 1800	4600	1600
Memory size, MB	1024/2048	1024/2048	1024/2048	1024/2048
	12	12	12	12
Interface	PCI-E3.0	PCI-E3.0	PCI-E3.0	PCI-E3.0
TDP level, W	65	65	30	30

For the Radeon R7 240, the core frequency is declared at 780 MHz. Given value often appears in all specifications, although it is Boost, and base frequency 730 MHz. For the Radeon R7 250, the base frequency is 1000 MHz with Boost 1050 MHz. In fact, most video cards operate on the Boost value, so you can operate with frequencies of 780 MHz and 1050 MHz in relation to voiced video cards. But there are models where the core frequency varies within the specified limits.

The Radeon R7 240 and Radeon R7 250 graphics cards are available in two versions - with DDR3 or GDDR5 memory. We will compare the two available models with DDR3, since these are the ones that currently dominate our market.

Let's start with the youngest video card made by Gigabyte. This model is equipped with 2 GB of DDR3 memory, and the letters OC in the code name indicate factory overclocking. Supplied in a compact box without additional accessories. The set includes a mandatory CD with software.

The Radeon R7 240 belongs to the category of the most affordable solutions on the market, but the manufacturer did not save on this model. It is made in a full-size version, and a large cooler is installed on top.

The total length of the Gigabyte GV-R724OC-2GI is 19.5 cm, including the protruding cooling shroud. The fan on the radiator is so large that it partially hangs over the textolite from the side.

Half of the memory chips are placed on the reverse side of the board. The fixing screws on the cooler do not have protective seals.

An aluminum radiator with a massive core and diverging petals is hidden under the plastic casing.

This design is often used in budget graphics cards. And this heatsink is much more solid than what we saw in the Gigabyte GV-N730D5OC-1GI.

The active cooling element is a 100 mm Eveflow T129215SM fan. Such giants are installed on more powerful video cards, it is unusual to watch it in such a product. With such a fan, you definitely do not have to worry about heat and noise.

Used textolite traditional for Gigabyte blue. The power supply of the core and memory has one phase each. It's nice to note the use of a high-quality element base, including Metal Choke chokes, which are also soldered in more expensive models.

Eight Nanya NT5CB128M16FP-EK memory chips are soldered.

And here is a photo of the Oland processor, which is quite modest in size:

The core frequency is increased from 780 MHz to 900 MHz, the memory runs on effective frequency 1600 MHz.

ASIQ Quality this instance rated at 72%.

Under gaming load, the video card did not warm up above 45-46 ° C at 23 ° C indoors. Below are screenshots of monitoring parameters during testing in Metro: Last Light and Tom Clancy's The Division. The fan spins up to 1200 rpm, the noise is minimal. Jumps in speed to large values ​​\u200b\u200bare not confirmed by real noise background, this is an inaccuracy in the sensor readings.

Overclocking potential is good, but with some limitations. For example, after exceeding the core frequency of 1100 MHz, a drop in memory frequencies was observed. And although the video card could exceed this level, we had to limit ourselves to 1100 MHz, especially since the memory frequency is already more important. The memory itself was able to function stably at 1980 MHz.

The resulting increase in core frequency was 41% relative to the recommended value of 780 MHz, the memory was overclocked by 24%. And such acceleration had practically no effect on heating, and the video card remained just as quiet.

The older Oland GPU comes in a familiar box. Adapters and additional cables are not supplied.

Externally, the older Gigabyte video card is completely identical to the younger one.

Gigabyte GV-R725OC-2GI also uses blue textolite and massive cooling with a 100 mm fan. Total length 19.5 cm.

When examining the reverse side, it is clear that other memory chips are used.

Gigabyte is equipped with three ports for connecting displays: HDMI, DVI and analog D-Sub.

The design of the cooling system is quite expectedly identical to that of the Radeon R7 240. The familiar aluminum radiator with divergent petals is used.

The radiator has a massive core. Radiators similar in design have been used in budget cards for a long time.

The fan Eveflow T129215SM with a size of 100 mm is responsible for blowing.

The printed circuit board is made according to a familiar scheme, but the power supply is amplified to three phases, two of which fall on the GPU power node.

The 2 GB memory is made up of eight Micron D9PRS chips.

The Oland XT processor uses a complete set of computing units.

For the Gigabyte GV-R725OC-2GI, the GPU frequency is increased to 1100 MHz, the memory runs at 1600 MHz.

The GPU-Z utility rated ASIQ Quality at 70.4%.

In terms of temperature and noise characteristics, the Radeon R7 250 does not differ from the younger model. Peak temperatures did not exceed 46-47°C, and the fan ran at less than 1200 rpm. In operation, such a video card is barely audible.

When trying to overclock the video card had to face some surprises. Any attempts to increase the memory frequencies did not lead to a practical result - the frequency of 1600 MHz remained unchanged during any operations. This is a sad fact, because memory frequencies can play the role of a deterrent. As for the overclocking of the GPU, the processor overcame 1205 MHz.

As a result, the video card was tested at frequencies of 1205/1600 MHz. It will be interesting to compare the results with the Radeon R7 240 at 1100/1980 MHz to see how important memory bandwidth is to unleash the potential of the Oland GPU.

Characteristics of the tested video cards

The reviewed video cards will be tested at their factory frequencies, at the recommended frequencies and overclocked. We will compare them with two versions of the GeForce GT 730 with different memory from the previous review. For competitors, we limited ourselves to standard frequencies, but added another participant from Gigabyte in the face of the GV-N730D5OC-1GI model with factory overclocking. The technical parameters of the participants are indicated in the table.

Video adapter		Radeon R7 250 DDR3		Radeon R7 240 DDR3	Gigabyte GV-N730D5OC-1GI	GeForce GT 730 GDDR5	GeForce GT 730 DDR3
Nucleus	Oland XT	Oland XT	Oland Pro	Oland Pro	GK208	GK208	GK208
Number of transistors, million pieces	1040	1040	1040	1040	1020	1020	1020
Process technology, nm	28	28	28	28	28	28	28
Core area, sq. mm	90	90	90	90	87	87	87
Number of stream processors	384	384	320	320	384	384	384
Number of texture blocks	24	24	20	20	32	32	32
Number of render units	8	8	8	8	8	8	8
Core frequency, MHz	1100	up to 1050	900	up to 780	1006	902	902
Memory bus, bit	128	128	128	128	64	64	64
Memory type	DDR3	DDR3	DDR3	DDR3	GDDR5	GDDR5	DDR3
Memory frequency, MHz	1600	1600	1600	1600	5012	5012	1600
Memory size, MB	2048	2048	2048	2048	1024	1024	1024
Supported version of DirectX	12	12	12	12	12	12	12
Interface	PCI-E3.0	PCI-E3.0	PCI-E3.0	PCI-E3.0	PCI-E 2.0	PCI-E 2.0	PCI-E 2.0
TDP level, W	65	65	30	30	25	25	23

test stand

Configuration test bench next:

• CPU: Intel Core i7-3930K (3, [email protected].4 GHz, 12 MB);
• cooler: Thermalright Venomous X;
• motherboard: ASUS Rampage IV Formula/Battlefield 3 (Intel X79 Express);
• memory: Kingston KHX2133C11D3K4/16GX (4x4 GB, [email protected] MHz, 10-11-10-28-1T);
• system disk: Intel SSD 520 Series 240GB (240 GB, SATA 6Gb/s);
• secondary drive: Hitachi HDS721010CLA332 (1 TB, SATA 3Gb/s, 7200 rpm);
• power supply: Seasonic SS-750KM (750 W);
• monitor: ASUS PB278Q (2560x1440, 27″);

• operating system: Windows 7 Ultimate SP1 x64;
• GeForce driver: NVIDIA GeForce 364.72;
• radeon driver: AMD Catalyst 16.4.1.

Testing was carried out at low graphics quality at a resolution of 1920x1080. If video cards did not provide an acceptable fps level in this mode, then a lower resolution of 1600x900 was set. Full description testing methods.

Test results

Battlefield 4

The game Battlefield 4 shows a clear dependence on memory frequency. With the same DDR3 memory, the difference between the Radeon R7 240 and Radeon R7 250 is less than 9%, and the younger video card is even faster in overclocking. The GeForce GT 730 with DDR3 is noticeably weaker than the Radeon R7 240. The GeForce GT 730 GDDR5 is already faster, but after increasing the frequencies, the Radeon R7 240 shows the same level of performance.

Dota 2

AMD representatives again occupy an intermediate position between the older and younger versions of the GeForce GT 730. The difference between the Radeons themselves is again small. The lack of memory overclocking limits the potential of the Radeon R7 250, while the Radeon R7 240 decisively takes the lead and catches up with the GeForce GT 730 GDDR5.

Fallout 4

The difference between Radeon R7 240 and Radeon R7 250 increases to 15% in Fallout 4. At the same time, the younger Radeon is faster than the younger rival from NVIDIA by 22-27%. Factory overclocking gives the Gigabyte GV-R724OC-2GI an additional 7% advantage. Gigabyte GV-R725OC-2GI, due to its frequencies, outperforms the simple version of the Radeon R7 250 by less than 2%. In overclocking, the one whose memory lends itself to acceleration is again better, despite the more modest core frequencies and fewer computing units in the R7 240. Overclocking also helps to minimize the gap between the Radeon R7 240 DDR3 and the GeForce GT 730 GDDR5.

Grand Theft Auto 5

In GTA, AMD participants were a little closer to the younger version of the GeForce GT 730, but still by a noticeable margin. At initial frequencies, the Radeon R7 240 DDR3 is 21-29% more productive than the GeForce GT 730 DDR3. The difference between the older and younger Radeon when compared at standard frequencies is 8-11%. Raising the frequency of a single GPU gives a meager performance change. Radeon R7 240, due to its overclocking of the processor and memory, brings the acceleration relative to the frequencies of 780/1600 MHz to 24-28%.

Metro: Last Light

AMD video adapters can hardly cope with Last Light at the minimum quality at 1600x900, but the increase in frequencies allows us to overcome the 30 fps mark. The Radeon R7 240 at recommended frequencies is 42% better than the younger GeForce, while the Radeon R7 250 has an additional 9% advantage. After overclocking, the younger Radeon easily outperforms its friend. The difference in results between frequency configurations 780/1600 MHz and 1100/1980 MHz is at the level of 25%.

Tom Clancy's The Division

Shooter The Division is even more difficult for budget participants. None of the tested video adapters will be able to comfortably play at the selected settings. So you will have to reduce the resolution to an unacceptably low level or think about buying more powerful models. If we talk about the difference between the participants, then the Radeon R7 240 with slow DDR3 memory is almost as good as the GeForce GT 730 GDDR5! And the Radeon R7 240 with frequencies of 1050/1600 MHz turns out to be slightly better than the factory overclocked Gigabyte GV-N730D5OC-1GI! It is noteworthy that for the first time we do not see the difference between the overclocked Radeon, they show the same results.

War Thunder

The Radeon R7 240 is once again closer to the GeForce GT 730 with fast memory than to the younger version of the GeForce. The backlog of the simple version of the Radeon R7 240 from such a competitor is 17-21%, and the Radeon R7 250 is already 1% faster than the GeForce GT 730 GDDR5. After overclocking, both Radeons achieve parity with the accelerated version of GeForce. It's nice to note that all AMD representatives provide comfortable high-resolution fps, if not at face value, then after overclocking.

3D Mark 11

In this test, the gap between the Radeon R7 240 and the junior GeForce is 8%. Factory frequencies provide Gigabyte GV-R724OC-2GI with an additional 10% advantage. There is a 31% difference between Radeon R7 240 and Radeon R7 250. The older Radeon is almost as good as the GeForce GT 730 GDDR5. After overclocking, the Radeon R7 250 maintains a slight advantage over the Radeon R7 240, falling slightly short of the level of the overclocked Gigabyte GT 730.

3D Mark Sky Diver

The positions of AMD video adapters are stronger here. The Radeon R7 250 is inherently better than the GeForce GT 730 GDDR5. The difference with the Radeon R7 240 at standard frequencies is up to 25%. After overclocking, both Radeons give the same results at the level of a senior NVIDIA representative.

3D Mark Fire Strike

The Radeon R7 240 DDR3 is 24% faster than the GeForce GT 730 DDR3, Gigabyte's model is another 7% faster. The difference with the Radeon R7 250 DDR3 is less than 20%. Overclocking brings the older Radeon minimal dividends. Radeon R7 240 at maximum frequencies is almost 30% faster than the initial 780/1600 MHz configuration. Maximum overclocking helps the Radeon R7 240 reach the performance level of the GeForce GT 730 GDDR5.

conclusions

The Radeon R7 240 and Radeon R7 250 graphics cards, combined with DDR3 memory, may well provide acceptable performance in games with low graphics quality. AT network games like Dota 2 and War Thunder, you can even count on medium quality settings at Full HD. But there are also games where the potential of these graphics cards is categorically lacking even at a lower working resolution. The most significant difference between these Radeon models in 3DMark tests, in real games, the advantage of the older card is small. To unlock the full potential of Oland GPUs, more fast memory. The example of the Radeon R7 250 with locked DDR3 frequency clearly shows that without increasing the memory bandwidth, overclocking the core does not make any practical sense. At the same time, after overclocking the video buffer, the Radeon R7 240 bypasses the older comrade, achieving a final advantage over its initial frequencies of 25-30%. Both models are noticeably faster than the GeForce GT 730 DDR3. It is harder to compete with the GeForce GT 730 GDDR5, but overclocking helps, which allows you to achieve close performance indicators. And this is a great result, because we are talking about comparing graphics solutions with a completely different type of memory.

The Gigabyte GV-R724OC-2GI video card is quiet and cold. Its high potential for further overclocking is quite realizable while maintaining the initial temperature and noise characteristics. And against the background of the fact that this model easily catches up with the more expensive Gigabyte card, it can be called the best purchase in case of a very limited budget.

Gigabyte GV-R725OC-2GI also boasts low low temperatures and low noise. All thanks to a powerful cooler that does a great job with more productive solutions. Increased core frequency and large quantity compute units give their advantages over the Radeon R7 240. It's frustrating that the memory cannot be overclocked and get additional acceleration, but it can be quite good. This problem may not be available on other units as the GV-R725OC-2GI is equipped with different DDR3 chips. Well, in general, this is not the first time that we are faced with the problem of memory overclocking on super-budget Radeons. For the money, this is a decent video card. And if you want overclocking, then, unfortunately, no one is immune from surprises.

The basis for assembling a computer is a video card. The final cost of the PC directly depends on its cost. For this reason, budget models are very popular on the video accelerator market. One of the brightest representatives is the AMD Radeon R7 240.

The accelerator appeared on the market in 2013. The video card can not boast of high performance. But, even despite this, its power is still enough to perform everyday tasks, as well as to run some modern games and programs.

The video card is based on the improved GCN architecture. According to user and expert feedback, the improvement had a positive impact on the performance of the accelerator. In addition, the R7 240 has added support for the API Mantle system.

AMD Oland PRO is used as the graphics core. Its frequency is 780 MHz in acceleration mode. The minimum frequency is 730 MHz.

The number of rasterization units of the device is 8. The amount of memory is 2 GB GDDR3 and 4 GB GDDR5. Due to the difference in memory size, the memory frequency readings are different. In the first case, it is 1600 MHz, and in the second - 4600 MHz. The throughput is 28.8 Gbps. The bus width is 128 bits.

The characteristics of the AMD Radeon R7 240 show that the graphics card will have enough power to work in office programs and applications, as well as for launching and comfortable game in projects with low system requirements.

Review Radeon R7 240

Two versions of AMD's video accelerator were released to the market: Radeon R7 240 4 Gb and R7 240 2 Gb. From these parameters, in addition to performance, the cost directly depends.

This model boasts an improved active cooling system. To do this, manufacturers have added an additional small fan to the existing heatsink. The only drawback of this solution was that the accelerator became noisier during the workflow.

The power consumption level of AMD Radeon R7 240 series is 50W. From this it follows that for the comfortable operation of the video card, a 300 W power supply is required. This indicator will be enough in conjunction with a productive processor and 6 GB of RAM.

To unleash the maximum potential of the video accelerator, you do not need to purchase a powerful processor. Therefore, a cheap AMD Phenom 2 X6 1055T model is suitable as a chip. This bundle will be more than enough for optimal performance. The device is equipped with three connectors: HDMI, VGA and DVI.

How to overclock AMD Radeon R7 240 graphics card

If you want to increase the basic power parameters, then you can overclock the AMD Radeon R7 240 video card. There are several ways to do this.

In the first case, you can use standard means Catalyst utilities. The disadvantage of this method is that the maximum allowable increase in GPU power through this utility cannot exceed 1000MHz. This limitation is set by the manufacturer.

To avoid such restrictions, you can use MSI program afterburner. With this utility you can increase the core speed up to 1100MHz.

To check the video accelerator for failures and problems, use the Funmark software. It allows you to conduct a stress test, at the end of which failures will be shown if they are detected.

To check the current R7 240 power ratings, download and run the GPU-Z utility. It will provide you with detailed information about the current technical parameters.

These overclocking parameters should be enough to solve more "heavy" tasks, such as video and image processing, running modern games.

Test results in games

To get a complete picture of the capabilities of a video card, you need to conduct tests in AMD games Radeon R7 240. Testing was carried out in the following projects.

Far Cry 3. The game was launched at high graphics settings in FullHD format. The FPS indicator was at the level of 35-37 frames. In serious action scenes and locations with a high density of objects: vegetation, NPCs, buildings, friezes and lags were not observed.

Alan Wake. On high settings at 1920×1080 resolution, the frame rate was 25-30. This is due to the fact that the game is not the most better optimization for computers. In places with a lot of lighting, microfreezes occurred.

Dota 2. MOBA game. In this project, the R7 240 proved to be excellent. Even at the highest graphics settings, the minimum FPS was a stable 35 frames, which is a very good value for a comfortable gameplay.

GTA 5. The launch was made on medium settings. FPS was kept at 25 frames. In some locations, there were drawdowns of up to 20 frames. This is due to the fact that GTA V is very demanding on the amount of video card memory. With a decrease in graphics quality, FPS rose to a stable 28.

Fallout 4. Even on the lowest graphics settings, the FPS in this game did not rise above 15 frames. This is not enough for a comfortable game. Also very often in the game there are lags. This is due to the fact that the locations do not have time to load to the end.

Dishonored. At high settings in FullHD format, the game showed stable and, most importantly, comfortable 27 frames. The minimum drawdown of this indicator occurs in moments with a large accumulation of NPCs.

Max Payne 3. The game is very well optimized for PC. Therefore, even at the highest graphics settings and FullHD format, the FPS indicator did not fall below 30 frames. No freezes or lags were found in the action scenes. The video accelerator handles Max Payne 3 very well.

Battlefield 4. Campaign Mode was at 20 FPS on Low graphics. When I increased the settings, it dropped to 15 FPS, ruining the entire gameplay experience. In online battles, the FPS is even lower - 17-18 frames with friezes in intense battles.

watch dogs. The game was launched at a resolution of 1920x1080 and low settings. The game is not distinguished by its optimization, and therefore the FPS indicator was at a rather low level - 23. Despite this, no serious lags were noticed during the gameplay.

wolfenstein. A game with very good optimization. Thanks to this, the FPS indicator was at a very comfortable level - 30 frames. The game settings were set to medium. Friezes and lags were not detected even during skirmishes with a large number of enemies.

Tomb Raider. When running the game on medium settings, the minimum frame rate was at an acceptable level of 25 FPS. This indicator is more than enough for playing in FullHD format.

Metro: Last Light. At medium graphic settings, the video card showed its best side. The minimum FPS was 27 frames at a resolution of 1920×1080 pixels. Freezes and lags during skirmishes with enemies were also not observed.

Based on the tests performed, we can conclude that the video accelerator can hardly cope with the launch of modern game projects. For the most part, the R7 240 is designed for office tasks rather than heavy applications.

Manufacturer comparison

Three major manufacturers are involved in the release of the video card on the market. To identify the most optimal, it is necessary to make a comparative analysis in the form of a table.

Manufacturer	HIS AMD Radeon R7 240	Asus Radeon R7 240	Gigabyte AMD Radeon R7 240
GPU	Oland PRO	Oland PRO	Oland PRO
Process technology	28 nm	28 nm	28 nm
Number of transistors	1040 million pieces	1040 million pieces	1040 million pieces
Render units	8	8	8
Crystal area (mm2)	90	90	90
Number of streaming multiprocessors	320	320	320
Video memory size (MB)	2048 and 4096	2048 and 4096	2048 and 4096
Video memory type	DDR3/GDDR5	DDR3/GDDR5	DDR3/GDDR5
GPU frequency (MHz)	780	730	900
CPU Temperature Limit (°C)	100	100	100
DirectX	12	12	12
Tire width	128 bit	128 bit	128 bit
Memory frequency	1600MHz DDR3/4600MHz GDDR5	1600MHz DDR3/4600MHz GDDR5	1600MHz DDR3/4600MHz GDDR5
Bandwidth (GB/s)	72	72	72
AMD Radeon R7 240 price, rub.	3999	4285	4598

There are no serious differences in the cost of the video accelerator. The most powerful in terms of technical parameters is the model from Gigabyte, as it has the highest core frequency.

Download drivers

Comes out every 2-3 months new driver for the Radeon R7 240 graphics card. Updating the software for the accelerator allows you to maintain its performance at an optimal level.

You can download drivers for the AMD Radeon R7 240 video card from the official website of the company.

Many months have passed since the release of AMD's current generation of video cards - the Radeon HD 7000 family. The first model of this line, the Radeon HD 7970, was announced almost two years ago! Since then, an updated version of the GHz Edition with an increased clock speed has been released, as well as a dual-chip Radeon HD 7990 and many models in other price categories, but we waited for a complete update of the line only today. True, the update turned out to be somewhat strange ... But let's not get ahead of ourselves.

AMD can fully consider the past two years to be successful. All graphics cards of this generation (Radeon HD 7900, HD 7800, HD 7700) sold well, and the Never Settle and Never Settle Forever programs, which involved the issuance of free coupons for the purchase of several popular games to buyers of AMD video cards, proved to be very successful, and increased volumes even more. sales of video cards of the company.

AMD is developing its approach to conquering markets, expanding its strategy. So, the company has invaded the field of game consoles even further (which we will discuss more than once below), not only offers video cards, but is seriously developing areas such as cloud computing, and helping producers of video games and other 3D applications in the development of content.

All this has certain consequences and to some extent changes the gaming market. Thus, the introduction of proprietary solutions (both CPU and GPU) in all next-generation game consoles that are about to enter the market has several consequences. For example, even purely theoretically, the development of multiplatform games should be seriously simplified, and the convergence of consoles and PCs in terms of hardware capabilities (both in functionality and performance) will give the much-awaited improvement in graphics quality and further strengthen the gaming PC market.

That's right: not only AMD and Nvidia consider the gaming PC market blooming and smelling. Many game developers, publishers and analysts vying with each other assure that PC games are the most alive and this market is only growing. Moreover, if you look at the diagram above, then analysts expect that already in 2013 the PC games market will surpass the console one, and in subsequent years, although it will yield slightly due to the release of new generation consoles, but even in such conditions it will be quite comparable with them. .

What does this mean for AMD and their competitors? That PC gamers will buy new games and upgrade their systems as the demands of future multi-platform titles increase dramatically. After all, the new generation of consoles have significantly increased capabilities compared to previous models. They have relatively powerful CPUs and GPUs, their memory capacity has increased by 16 times, and they are comparable, if not with top-end PC solutions, but with systems in the upper-mid price range. And given the fact that hardware consoles are traditionally squeezed more than PCs, it can be assumed that new games will significantly raise the bar for system requirements.

Moreover, the PC is always ahead of the consoles, in particular in support of high-resolution display devices. For example, Ultra HD (“4K”) resolution displays are already on the market, requiring four times the power from GPUs compared to currently common Full HD systems. And while these monitors are still quite rare, they are expected to enter the market very soon, and the price reduction should serve them well. Gradually, a new era of PC gaming is dawning, with the influence of Ultra HD resolution and new generation consoles, where many graphics cards in gamers' systems will need to be upgraded.

That's why AMD today announced a new generation of its Radeon graphics cards. More precisely, it now contains several series: the R9 and R7 series (in the future, the budget R5 series is also expected, but for the players it is simply not interesting, since it rather plays on the APU field). The dual new line of the company contains the following models, covering most of the market segments:

Thus, video cards of R7 250 and R7 260X models are designed for the price range of $90-$140 (prices in the US market), R9 270X will be sold for $200, and R9 280X - for $300. Unfortunately, there will be no detailed information about the flagship of the line - the R9 290X model - today, the announcement of this model will take place separately.

But it is already known that the company offers to purchase an exclusive edition of AMD Radeon R9 290X Battlefield 4 Edition. As the name suggests, this video card will include the game of the same name, which is coming out this month. This edition will be released in a limited edition, and Battlefield 4 will not be included with other graphics cards, so the package is truly unique.

Material about video cards from the AMD Radeon R9 290 series will be released later, but for now we can tell you about this line that it will be based on a completely new GPU codenamed Hawaii (the current generation top chip is codenamed Tahiti), highly energy efficient, based on the improved Graphics Core Next architecture and supporting the latest version of the DirectX 11.2 graphics API.

The new top-end graphics card of the R9 series will have a peak math performance of more than 5 teraflops, more than 300 GB / s of video memory bandwidth, it is capable of processing more than 4 billion polygons per second. Therefore, it is not surprising that the Hawaii chip is much more complex than Tahiti and consists of more than 6 billion transistors. You will know the exact numbers very soon, but for now let's look at all the other models of the updated line of AMD video cards.

Since the new Radeon R7 and R9 solutions largely repeat the features of the previous Radeon HD 7000 series, it will be useful to familiarize yourself with detailed information about early AMD solutions:

• AMD Radeon HD 7870: mid-range 3D graphics solution based on GCN architecture
• AMD Radeon HD 7770/7750: New Architecture Goes Mainstream
• AMD Radeon HD 7970: New Single-GPU 3D Graphics Leader

Let's move on to the description of the technical characteristics of the announced video cards of the new family.

AMD Radeon R7 and R9 family graphics cards

Let's take a closer look at all the new products from AMD. First, a few words about new system items. In our opinion, it is not ideal, although it is somewhat similar to the one that has long been used in APUs (A8 and A10 families, for example) and other manufacturers (for example, Core i5 and i7). And yet, for video cards, the previous naming system was clearer, and it’s surprising that AMD decided to change it right now, although they had at least the Radeon HD 9000 line in stock. And the “HD” prefix could simply be changed to something another (yes, at least "UHD" - from Ultra HD!). The division into the R7 and R9 families also remains unclear: why does the 260X still belong to the R7 family, while the 270X already belongs to the R9?

But let's leave the dispute about the names - after all, they do not affect anything, unlike the technical characteristics, which we will now consider. On the one hand, this section of the article is the most important: it will provide technical specifications and give preliminary estimate performance of new solutions. On the other hand, in practice it turned out that there are only two completely new solutions in the R7 and R9 lines - R9 290 and R9 290X, and we are not ready to talk about them yet.

How is it that there are almost no new solutions among all these video cards? The point here is that although these models are nominally new, almost all of them are based on the same GPUs known to us from the previous Radeon HD 7000 line. that this is a slightly modified Radeon HD 7970 GHz Edition: it is based on the same Tahiti video chip and has the same key characteristics.

The same applies to some other solutions of the new series, although not all. For example, the Radeon R9 270X is based on a new chip with the code name Curacao, but what is its difference from Pitcairn and why the release of a new, but almost the same chip was needed is not clear. The Radeon R7 260X is based on the Bonaire chip, known from the Radeon HD 7790, while the younger R7 240 and R7 250 solutions are based on the Oland GPU, which has not yet been used in desktop video cards. However, there is also nothing particularly interesting in it, and the number of functional blocks in this budget GPU is even less than in Cape Verde. But let's look at the characteristics of the new line:

AMD Radeon R9 280X graphics card

• Chip codename: "Tahiti"
• Core frequency: up to 1000 MHz
• Number of universal processors: 2048
• Number of texture units: 128, blending units: 32
• Effective memory frequency: 6000 MHz (4×1500 MHz)
• Memory type: GDDR5
• Memory bus: 384 bits
• Memory capacity: 3 gigabytes
• Memory bandwidth: 288 gigabytes per second
• Compute performance (FP32): 4.1 teraflops
• Theoretical maximum fill rate: 32.0 gigapixels per second.
• Theoretical texture sampling rate: 128.0 gigatexels per second
• Two CrossFire connectors
• PCI Express 3.0 bus
• Power consumption: 3 to 250 W
• One 8-pin and one 6-pin power connectors
• Dual slot design
• US MSRP: $299

This model is located in the new line of the company one step below the top R9 290(X), which has not yet been released completely. It is based on the successful Tahiti video chip, which was the top one recently, and is a complete analog of the Radeon HD 7970 GHz model, but it is already sold for $299 (on the US market). Among the advantages of the model, AMD calls the amount of video memory of 3 gigabytes, which will be in demand at high resolutions, such as 2560 × 1440 and Ultra HD, in demanding games such as Battlefield 4. Moreover, the amount of video memory of 3 GB is the official recommendation of the developers of this game. .

As for comparing performance and price with previous solutions, then, following a competitor, AMD fell in love with comparisons with video cards from many years ago. Of course, the new product will look just fine if you compare it with the Radeon HD 5870, which came out ... already 4 years ago:

The graphics cards in the chart are compared in a modern 3DMark test suite, so it's not surprising that the R9 280X is more than twice as fast as the top-end motherboard from years ago. More importantly, this performance is offered for around $300, which is pretty good, although some Radeon HD 7970 models are already selling for nearly the same amount.

If we compare it with the competitor's solutions, then AMD claims an average advantage of 20-25% over the Geforce GTX 760 video card of the competing Nvidia, which has a similar price. This is probably true somewhere, we will check this in future practice materials, the first of which will appear at the end of the month.

AMD Radeon R9 270X graphics card

• Chip codename: "Curacao"
• Core frequency: up to 1050 MHz
• Number of universal processors: 1280
• Number of texture units: 80, blending units: 32
• Effective memory frequency: 5600 MHz (4×1400 MHz)
• Memory type: GDDR5
• Memory bus: 256 bits
• Memory capacity: 2 or 4 gigabytes
• Memory bandwidth: 179 gigabytes per second
• Compute performance (FP32): 2.7 teraflops
• Theoretical maximum fill rate: 33.6 gigapixels per second.
• Theoretical texture sampling rate: 84.0 gigatexels per second
• One CrossFire connector
• PCI Express 3.0 bus
• Connectors: two DVI Dual Link, HDMI 1.4, DisplayPort 1.2
• Power consumption: 3 to 180 W
• Two 6-pin power connectors
• Dual slot design
• US MSRP: $199 (4GB model $229)

The R9 270X sits in the middle of AMD's Radeon lineup and is based on the new Curacao video chip, which is practically the twin of Pitcairn. This video card almost completely repeats the Radeon HD 7870 model known from the previous line, but it will be sold in the North American market for only $199, although it also differs from last year's card in terms of speed, and they consist in increased clock frequency GPU and video memory, which should have a positive impact on performance. Moreover, the maximum frequencies themselves mean little now - in practice, the GPU can operate at an even higher frequency, and the R9 270X will be closer in speed to the Radeon HD 7950 than to the HD 7870.

The model under consideration has a video memory capacity of two gigabytes, which is quite enough for resolutions up to 1920×1080(1200) even in modern demanding games at high settings. Traditionally, the performance and price of new products are compared with previous solutions. This time, for comparison, we also took a four-year-old Radeon HD 5850 model, which at one time even had a slightly higher price:

Not surprisingly, the Radeon R9 270X delivers more than double the performance in modern benchmarks compared to one of the older models. And the second one - Radeon HD 6870 - is ahead by almost the same margin. As for the comparison with Nvidia video cards, AMD compares the new product with the Geforce GTX 660 model, believing that its $199 version is 25-40% faster than its competitor in a specially selected set of modern games.

AMD Radeon R7 260X graphics card

• Chip codename: "Bonaire"
• Core frequency: up to 1100 MHz
• Number of universal processors: 896
• Number of texture units: 56, blending units: 16
• Effective memory frequency: 6500 MHz (4×1625 MHz)
• Memory type: GDDR5
• Memory bus: 128 bits
• Memory capacity: 2 gigabytes
• Memory bandwidth: 104 gigabytes per second
• Compute performance (FP32): 2.0 teraflops
• Theoretical maximum fill rate: 17.6 gigapixels per second.
• Theoretical texture sampling rate: 61.6 gigatexels per second.
• One CrossFire connector
• PCI Express 3.0 bus
• Connectors: two DVI Dual Link, HDMI 1.4, DisplayPort 1.2
• Power consumption: 3 to 115 W
• One 6-pin power connector
• Dual slot design
• US MSRP: $139

The third model unveiled today is priced even lower at $139 and is an almost complete copy of the Radeon HD 7790 and is based on the same GPU, codenamed Bonaire. Among the differences between the new model and the old one from the previous line is a slightly increased frequency and the presence of two gigabytes of video memory. This is understandable: memory requirements are growing very quickly over time, and this will be even more evident with the release of multi-platform games designed for next-generation consoles.

The Radeon R7 260X has enough performance for undemanding gamers, enough for high settings quality in most games. AMD compares the performance and price of the novelty with only one of the previous generations of video cards - the Radeon HD 5870, again four years ago:

Apparently, the outdated top board was taken in order to show that the performance of the former representatives of the high-end segment is now available for only $139 (again, all prices are in the US market), and the novelty even has a power reserve. Of the competing solutions, AMD mentions Nvidia model Geforce GTX 650 Ti, and on the diagrams of this company, the new model R7 260X is 15-25% faster than its rival.

AMD Radeon R7 250 graphics card

• Chip codename: "Oland XT"
• Core frequency: up to 1050 MHz
• Number of universal processors: 384
• Number of texture units: 24, blending units: 8
• Effective memory frequency: 4600 MHz (4×1150 MHz)
• Memory type: GDDR5 or DDR3
• Memory bus: 128 bits
• Memory bandwidth: 74 gigabytes per second
• Compute performance (FP32): 0.8 teraflops
• Theoretical maximum fill rate: 8.4 gigapixels per second.
• Theoretical texture sampling rate: 25.2 gigatexels per second
• PCI Express 3.0 bus
• Connectors: DVI Dual Link, HDMI 1.4, VGA
• Power consumption: 3 to 65 W
• Dual slot design
• US MSRP: $89

Perhaps this is the first video card from the new AMD line that does not have a clear predecessor in the retail line, since the Oland chip is used in desktop solutions for the first time (it was used in OEM solutions of the Radeon HD 8000 family, which is not very well known to the general public). This is the most affordable graphics card based on the Graphics Core Next architecture GPU, designed for the entry-level price segment - it costs less than $90!

Radeon R7 250 video cards will be available both in two-slot and single-slot versions, depending on the manufacturer's decision. Naturally, such a video card does not need additional power - it is content with the energy received via PCI-E. Let's see what it has to offer in terms of performance:

And again, AMD compares the latest model with a solution from the distant Radeon HD 5000 family. Now the mid-range video card is taken - HD 5770, which at one time had considerable success on the market. So, the current budget model provides performance higher than the old one, and this at almost half the price! By now, this is the most entry level for modern 3D games, and below it in performance - only the APU and ... one more new graphics card family R7.

AMD Radeon R7 240 graphics card

• Chip codename: "Oland Pro"
• Core frequency: up to 780 MHz
• Number of universal processors: 320
• Number of texture units: 20, blending units: 8
• Effective memory frequency: 4600 MHz (4×1150 MHz) or 1800 MHz (2×900 MHz)
• Memory type: GDDR5 or DDR3
• Memory bus: 128 bits
• Memory capacity: 1 (GDDR5) or 2 gigabytes (DDR3)
• Memory bandwidth: 74 (GDDR5) or 23 (DDR3) gigabytes per second
• Compute performance (FP32): 0.5 teraflops
• Theoretical maximum fill rate: 6.2 gigapixels per second.
• Theoretical texture sampling rate: 15.6 gigatexels per second.
• PCI Express 3.0 bus
• Power consumption: 3 to 30 W
• Single slot design

In fact, this is an even cheaper version of a video card based on the Oland video chip. It has a slightly truncated GPU running at lower frequencies, and it is likely that most of these graphics cards on the market will have slow DDR3 memory, which will affect their 3D performance. However, for such cheap motherboards, performance is no longer important. Moreover, even less expensive solutions of the R5 family may appear in the future, but this is a different story.

It's no wonder that AMD's partners are ready to supply solutions of new families almost from the moment of announcement, and even with their own design of boards, coolers and factory overclocking. Indeed, for many of the new products, they just need to flash slightly modified BIOS versions, change the design of boxes and coolers - and here are the new products:

Actually, even practical tests in games on new video cards are not very interesting - you can simply take as a basis the results of those video cards of the past generation, almost complete copies of which are models from new families, and add 5-15% of the advantage obtained due to increased frequencies and tweaked power management technologies. After all, only the R7 240 and R7 250 have obvious differences from the boards of the Radeon HD 7000 family, and the rest of the cards (well, except for the R9 290 and 290X, which have not yet been released) are renamed old boards. And when there are no hardware changes, there is usually a lot of talk about new software technologies, which let's move on to.

Mantle - low-level graphics API

Perhaps the most unexpected announcement, along with the new line of AMD Radeon graphics cards, was the introduction of a new graphics API, dubbed Mantle. AMD, despite good relations with the Microsoft DirectX development team and support for the latest version of this API (DirectX 11.2) by its video chips, decided to take such a serious step. Of course, they were inspired by the fact that in the next generation of game consoles, it is AMD that will be the supplier of absolutely all GPUs for all companies: Sony, Microsoft and Nintendo, and from this you can at least try to have some advantage.

It seems that AMD decided to release such an API largely due to the influence of DICE and EA, which released the Frostbite game engine that underpins Battlefield. To understand what Mantle is and why it is needed, it is necessary to bring the point of view of one of the leading game developers. The AMD event featured a presentation by Johan Andersson, the CTO at DICE, who is responsible for the Frostbite engine. He said that they consider the PC to be a great gaming platform with rich features, and moreover, for DICE, the PC has been the main platform since Battlefield 1942, and they promise to support PC games and beyond.

AMD and DICE have been working together for a long time - it all started with Battlefield 2 in 2004. Collaborations between the two companies include new technology development teams working together, the introduction of technologies such as Eyefinity and CrossFire, and more, such as a special 4K demo of Battlefield 4 on dual Radeon HD 7990 graphics cards at the GDC Game Developers Conference.

Frostbite 3 is a new DICE engine, and at the same time it is a platform for many other EA games: shooters, strategies, RPGs, racing, and more. this moment more than 15 games in the Battlefield, Need for Speed, Star Wars series are in development, mass effect, Command & Conquer, Dragon Age, Mirror's Edge, and more, making Frostbite optimization for AMD GPUs a top priority.

This engine is very modern, it uses "native" 64-bit executable code with the ability to run 32-bit on older CPUs, the capabilities of eight CPU cores are used, the engine is optimized for AMD Radeon video cards and DirectX 11.1 - it is in this version of the graphics API by Microsoft at the request of DICE, certain graphics capabilities. Here are just some of the features of Frostbite using Battlefield 4 as an example: level destructibility, imitation of a water surface for multiplayer, complex visual effects, lighting using compute shaders, complex post-filtering: DOF with bokeh, subsurface scattering, motion blur, supersampling.

But when creating PC versions of multi-platform projects, there are always some difficulties. While the Frostbite engine scales well from low-end to high-end systems, all hardware configurations need to be supported, providing a wide range of graphic settings. Also on PC, it is not possible to use all CPU cores in the game's graphics engine due to DirectX and OpenGL limitations, and extra costs CPU resources in these APIs hinder development and slow down code.

And some features that are available on the PC are simply impossible to open due to existing restrictions that appeared many years ago. Initially, on the PC, it was so that the CPU "feeds" the data to the GPU, and the close interaction between them when working on the same tasks is very limited. At the same time, consoles have been doing it for a long time so that part of the work (for example, post-filtering) is done on the CPU, and part on the GPU, and their memory access is equally or almost equally fast.

Also, not all hardware capabilities of the released GPUs can be used with existing graphics APIs. Some functionality that exceeds the DirectX and OpenGL specifications remains unused by developers. The slow development of graphics APIs does not suit everyone, and some of the developers want to use all the hardware capabilities, not being limited by the current software limits and using a "thinner" shell between the game engine and GPU hardware resources.

Consoles do not have all these problems, since they have a single fixed hardware and software configuration, almost all of whose features are available in the development of games and applications. Also, operating systems and APIs on consoles are a much less thin layer between applications and hardware, allowing for simplified development and low-level access to many advanced features.

Considering that all future game consoles of the “desktop” format (Playstation 4 and Xbox One, first of all) are based on AMD graphics solutions based on the GCN architecture familiar from PCs, AMD and game developers have an interesting opportunity to take advantage of this to their advantage by releasing a dedicated graphics API that will allow game engines to be programmed on PC in the same style as on consoles, with minimal API impact on the game engine code. The same DICE has long dreamed of a similar approach and talked with GPU manufacturers, and now such an opportunity has appeared.

Mantle is a low-level, high-performance "console-style" graphics API for the PC that was developed at AMD with significant input from top game developers like DICE. This is not surprising: DICE develops, and EA releases multi-platform games, in facilitating development and improving functionality which they are interested. Battlefield 4 is the first project to use Mantle, all other developers will be able to use this API in the future.

According to preliminary data, the use of Mantle provides a ninefold advantage in the possible number of draw calls (draw calls) compared to other graphics APIs, which reduces the load on the CPU. Of course, such a multiple advantage is possible only in artificial conditions, but some superiority will be provided in typical 3D game conditions; the question is what. In any case, the announcement of Mantle is a very high-profile event in the world of PC graphics, which can give an additional impetus to the development of new graphics algorithms and techniques, facilitate their transfer from consoles to PC and vice versa, and also strengthen the development of multiplatform game engines.

Although Battlefield 4 will be released at the end of October, the release version will only support DirectX 11.1, and Mantle API support is scheduled for December, when a special free update will be released, further optimized for AMD Radeon graphics cards. On PC systems with GCN architecture graphics cards, the Frostbite 3 engine will use Mantle, which will reduce the load on the CPU, parallelize work across eight processing cores, which is not possible in the regular version, introduce special low-level performance optimizations and full access to GCN hardware capabilities. And this is just the beginning - other interesting ideas are possible in the future, such as entire worlds "living" on the GPU, or low-level rendering on multiple GPUs that does not use CrossFire at all.

To all questions about the practical differences between the Mantle and DirectX versions of Battlefield 4, and at least the approximate performance increase that is expected, AMD representatives answered in silence. Apparently, this is due to the fact that the work of DICE has not yet been completed and so far there are not even approximate figures from them. Moreover, Mantle still has more questions than answers. How will the low-level Mantle driver work with its direct access to GPU resources in operating system Windows with DirectX managing GPU resources themselves? How will these resources be shared between the Mantle game application and the Windows system?

Answers to these questions and more are expected no earlier than mid-November 2013, when the AMD Developer Summit will be held, which will reveal the technical details of the implementation of Mantle, the list of partners, and even show demo programs. We really hope to get all the information you are interested in, as well as to find out about those developers who are interested in this API, as this is something new in 3D graphics on the PC. Something that could theoretically change the industry. Or maybe not if, for example, manufacturers of game engines and games decide that it will be too expensive for them to develop two directions at once (DirectX and Mantle).

TrueAudio sound processing technology

Another unexpected and curious announcement from AMD was the technology associated with ... sound. In general, AMD has always paid a lot of attention to sound. In 2006, they first released solutions capable of transmitting audio data over an HDMI cable directly from ATI Radeon HD 2000 series video cards, in 2008 they made support for DisplayPort audio in the ATI Radeon HD 3600, in 2009 - support for high-frequency audio transmission. bitrate over HDMI in the Radeon HD 5800 series, and so on.

But all this was not connected with the actual sound processing. With the release of the Radeon R7 and R9 series, the company introduced AMD TrueAudio technology to the world, a programmable audio engine that appeared in some of the graphics cards released as part of the new series. Yes, unfortunately, TrueAudio is only supported on AMD Radeon R7 260X and unannounced top R9 series solutions. This is understandable: only Bonaire and Hawaii chips are the newest in terms of technology, they have the GCN 1.1 architecture and other innovations, including TrueAudio support. And this is one of the most important limitations.

What is TrueAudio? On the PC, and in games in particular, support for hardware sound processing has long been consigned to oblivion. First, such giants as Aureal were absorbed (a very long time ago), then the positions of the even larger giant Creative were pretty shaken under the onslaught of built-in motherboards audio codecs with primitive capabilities, and finally finished off hardware sound on a Microsoft PC by disabling support for DirectSound and DirectSound3D hardware acceleration in the operating Windows system Vista.

In contrast, audio processing on gaming consoles has always been handled by specialized hardware units. As a result, lately the PC has been inferior to them in terms of sound quality, and in the corresponding versions of multi-platform games, we do not hear what console players hear. What explains this? The fact that universal CPU cores are far from ideal for sound processing, and they are engaged in many other tasks, among other things. The budget of processor time allocated in games to sound is not that big (taking into account the versatility of CPU cores), and some effects have to be sacrificed during software processing.

As you can see, in this sample game, 10% of the processing resources of the available CPU are allocated to sound processing. This is not always enough. So AMD decided to go the "console" route by embedding a fully programmable audio engine into its own GPUs, the first of which was the Bonaire chip on which the Radeon R7 260X is based. TrueAudio technology gives developers the flexibility and high performance they need to process audio with different algorithms, here is a partial list: more mixing sounds, sound level equalization, complex reverb, etc.

TrueAudio provides guaranteed real-time processing of audio tasks on a system with a supported GPU, regardless of the installed CPU. To do this, several Tensilica HiFi EP Audio DSP DSP cores were integrated into Hawaii and Bonaire chips, the capabilities of which can be read. However, TrueAudio hardware is not limited to DSP cores, here is a detailed slide with the architecture of the hardware that is included in some of the new GPUs:

The diagram shows several Tensilica HiFi 2 EP DSP cores optimized for sound processing, Tensilica Xtensa floating point data processors, as well as caches and internal memory (32 KB cache for data and instructions and 8 KB local "scratch" - memory per DSP), multi-channel DMA engine, built-in 384 KB shared memory, access interface to system memory, up to 64 MB of addressable space in video memory, etc.

TrueAudio's power is accessed through popular audio processing libraries used by game developers, and the technology is completely changing the way we sound games. Sound engine and effects developers can use the resources of the built-in audio engine using the dedicated AMD TrueAudio API.

Naturally, in the case of any new technologies, the issue of partnership with the developers of audio engines and libraries for working with sound is very important. And AMD is trying to work closely with many companies known for their developments in this area. At the presentation of the new products of the Radeon R7 and R9 families, several representatives of AMD's audio processing partners spoke about the inclusion of TrueAudio support in their future applications and games.

The list of partners is quite good, it includes game developers (Eidos Interactive, Creative Assembly, Xaviant, Airtight Games), and audio middleware developers (FMOD, Audiokinetic), and audio algorithm developers (GenAudio, McDSP), and this is just the beginning. A GenAudio representative spoke about the AstoundSound technology, which allows you to position sounds in a spherical space around the user not only horizontally, but it does a pretty good job even with a vertical direction.

AstoundSound technology is available as plugins for the popular FMOD and Wwise sound engines and is fairly easy to integrate into games. Support for AMD TrueAudio helps offload the CPU, increase the number of simultaneously processed sounds and is multi-platform, as the consoles also have dedicated DSPs for sound processing.

One of interesting opportunities that game developers plan to use in their projects is convolution reverb - reverb based on digital convolution of the processed sound signal with impulse response (IR). Simply put, this reverb uses "recordings" of real rooms - as if the sound image of the room, expressed in mathematical form.

The convolution reverb process simulates the reverberation of a real physical space based on a pre-recorded "record" (impulse response) of this simulated space. The advantage of this approach compared to the reverb presets we have seen in EAX, for example, is that reverb based on digital audio convolution provides a realistic reproduction of sound in rooms and outdoors, without being limited by a predetermined quantity and quality. presets.

But this algorithm is difficult to programmatically execute on the CPU, since it is very demanding on computing power (10-15% of CPU resources can be easily taken up), and it also needs quite active work with memory during processing. TrueAudio technology provides reverb based on the digital convolution of the audio signal, almost completely freeing the CPU from this difficult task. For games, this means that more complex algorithms can be used with TrueAudio.

By the way, Eidos' Thief game from the well-known game series for many years, which is a simulator of a thief with a first-person view, planned for release early next year, plans to introduce AMD TrueAudio sound technology. This is not surprising, because it is in such games, where the gameplay depends on high-quality positioned and simulated sound almost more than on the visual part, that good sound is needed.

All in all, the TrueAudio technology is quite interesting, especially given the apparent stagnation of hardware audio processing on the PC and its active use on consoles. The question, as always, is the relevance of the solution at the moment. How many game developers will rush to integrate the technology into their projects, given that at the moment it is available only on one video card (Radeon R7 260X)? Yes, R9 290 series boards will appear eventually, and all the next AMD GPUs will contain dedicated audio DSPs, so TrueAudio can become really in demand. Whether it will actually become such - only time will answer this question. In any case, innovations in the field of sound can only be welcomed, otherwise this swamp has stagnated too much.

Display technologies: support for Ultra HD and Eyefinity

AMD has long been one of the leaders among the companies that are pioneers in the field of information output to display devices: monitors, TVs, projectors ... For example, AMD was the first or one of the first among those who introduced DVI Dual Link support for monitors with a resolution of 2560 × 1600 pixels (October 2005), DisplayPort support (January 2008), output to three or more monitors - Eyefinity technology (September 2009), and then this support improved - in October 2011, portrait multi-monitor was supported 5×1 mode, etc. In December 2011, AMD was the first to introduce support for DisplayPort 1.2, and in February 2012, HDMI with 4K resolution.

4K resolution, also known as Ultra HD, is 3840x2160 pixels, exactly four times the resolution of Full HD (1920x1080), and is very important to the industry. It is 4K that can give another serious impetus to the development of all companies associated with images - after all, users have also been waiting for something really new, and here they will have a fourfold improvement in detail.

The only problem is the low prevalence of Ultra HD monitors and TVs at the present time. 4K TVs are sold only very large and expensive, and the corresponding monitors are extremely rare (the number of models can be counted on the fingers) and also super expensive. But the situation is about to change if the forecasts of analysts predicting a bright future for Ultra HD devices come true:

AMD provides connectivity for two options Ultra HD displays: TVs that only support 30Hz and below at 3840x2160 resolution and connect via HDMI or DisplayPort, as well as monitors that are split in half at 1920x2160 resolution at 60Hz. The second type of monitors is also supported with DisplayPort 1.2 MST hubs, which have recently gone on sale.

In general, with support for tiled 4K displays, in reality, everything is not so simple. To support such a high resolution at 60 Hz, two video streams are required, as one cannot provide the required bandwidth. After all, if to transmit an image with HD resolution, a bandwidth of less than 100 MP per second is required, then Full HD resolution requires about 140 MP/s, and Ultra HD requires more than 500 MP/s! Therefore, such displays are supported by the previous AMD Radeon HD 7000 line when using two video outputs or MST streams through special DisplayPort hubs.

To support split monitors was introduced new standard VESA Display ID 1.3, which describes additional display features such as identifying tiled devices, describing the tiling topology and positioning of each tile, attaching a specific stream to a single tile, and describing the position and size of frames. All this will make it easier to configure complex multi-monitor designs created using AMD Eyefinity technology, because with all this data, setup will be much easier.

The new VESA standard will automatically "glue" the image for such monitors, if supported by both the monitor and the driver. This is planned for the future, but for now, these 4K tiled monitors require manual configuration. AMD says that in latest versions Catalyst drivers there is already an automatic configuration option for the most popular monitor models.

By the way, about the future of Ultra HD monitors. The following models of AMD Radeon graphics cards will support the third type of Ultra HD display, which requires only one thread to run at ultra-high resolution at a refresh rate of 60 Hz. Planned AMD Graphics Card Models Ready to Support high speeds data transfers up to 600 MHz, which is required for this, and we can only wait for the appearance of new Radeons and monitors with corresponding support.

This section would not be complete without new information about AMD Eyefinity technology. It is well known that the AMD Radeon HD 7000 series and earlier families currently support up to two HDMI/DVI displays, and all other devices in a multi-monitor configuration must have a DisplayPort input or connect using active DisplayPort dongles.

The AMD Radeon R9 Series already supports up to three HDMI/DVI displays with AMD Eyefinity technology. This feature requires a set of three identical displays that support identical timings, output is configured at system startup, and does not support display hot-plugging for a third HDMI/DVI connection. At the same time, DisplayPort connectors can be used, which will increase the number of monitors supported by one video card to six.

Software support: Raptr and the new Ruby

We have already mentioned that AMD continues to improve software support for its solutions. So, together with Raptr, specialized software was created that is designed to make life easier for the gaming community. This software is designed to solve several problems at once that arise in front of a PC player. PC games are a very good thing, they are always technically perfect and have enough options for customization to the user's requirements, but this also has its drawbacks. Not all players want to fiddle with the settings for a long time, adjusting the games for themselves, some just want to press a button and play.

But on a PC there is no such possibility due to the many software and hardware configurations, and there are practically no services like the console Xbox Live. AMD's main competitor, Nvidia, released software some time ago that makes it easy to at least adjust the graphics settings in games, which makes setting up and running PC games close to what is available on consoles - just press one button, and the application will be optimized for specific system.

In the case of AMD, such an application is called Raptr, it is already possible, but it is not limited to the specified functionality. This utility has collected a lot of features demanded by the gaming community into one heap, and at the same time it is not limited to any individual publishers or platforms, but is a unifying tool for all players. By the way, according to the company, there are already more than 18 million players in the Raptr community - this is a very impressive figure.

Other features of Raptr include access to your favorite applications directly from games without having to switch between windows, the ability to broadcast video of the gameplay to everyone, as well as various additions typical of gaming communities: rewards for time spent in games; free games and add-ons, betas and discounts on full versions applications.

Nevertheless, the main thing for us is the ability to determine the optimal game settings for a particular game. game system from the CPU and GPU installed in the PC. This Raptr functionality is easy to use. The software at startup determines Hardware, finds installed games, as well as their settings, then builds FPS graphs during the game and searches for optimal settings. In addition, Raptr uses real frame rate data from other users on similar systems.

As with the corresponding Nvidia software, optimization requires only a click of the mouse, but unlike the Geforce Experience, there are three possible settings here: Performance, Balanced and Quality. Another important difference from GFE is the use of data not from the test lab, but from all users who have ever launched the game at different settings- Raptr collects all this data and automatically finds optimal settings based on the mass of the analyzed material. However, it looks smooth on paper, but we'll see how it will be in reality.

And finally, let's talk about the pleasant. For 10 years now, AMD has been creating and showing demo programs, the main character of which is a girl named Ruby. In the latest version, developed for the current announcement by AMD, Illfonic and Crytek, it has seriously changed its appearance - it has undergone a clear restyling.

This demo uses CryEngine and is optimized for Graphics Core Next architecture GPUs. The demo uses 17 technologies made possible by DirectX 11 support, including the well-known TressFX physical hair simulation, which simulates 12,000 individual Ruby hairs. It seems that it is the need to show the capabilities of TressFX that explains the change in the girl's appearance - after all, she previously wore a short haircut.

conclusions

Although we will draw final conclusions on the line of AMD Radeon R7 and R9 video cards presented today after practical testing, which is expected on our website towards the end of October, we still allow ourselves to express some considerations. Despite the fact that many of the announced models of the new line are simply renamed and slightly overclocked versions of the already known models of the Radeon HD 7000 family, the release of new products can be generally assessed positively, and here's why.

First, AMD offers very competitive prices for the entire line, from the R9 280X, formerly known as the Radeon HD 7970 GHz Edition, to the budget R7 series boards based on the new Oland video chip. With such prices, almost all presented models of AMD video cards from new families have a very good ratio of price, performance and functionality.

Secondly, the same functionality is only expanding and improving. Along with the announcement of new solutions from the Radeon R7 and R9 families, we also got acquainted with very interesting initiatives from AMD: the GPU-integrated audio DSP engine in the form of TrueAudio technology and the new Mantle graphics API, the development and announcement of which became possible largely due to the fact that AMD has won the role of graphics solution provider for all next-generation game consoles.

Yes, so far, the prospects for these exciting initiatives in PC games are very vague, and it is far from certain that they will become widespread among game developers, despite all their advantages and innovation. We will see with interest how AMD will succeed in promoting its technologies, since this is a very difficult and time-consuming task. After all, creating your own standards is much more difficult than simply using those recognized by the entire industry ...

And the last thing: it seems that AMD has not yet said its most important word in the form of the release of the top products of the line, known as the Radeon R9 290(X). It is these solutions, based on the latest top-end graphics processor codenamed Hawaii, that should become the locomotive that will pull along all new technologies (Mantle and TrueAudio) and the entire modern product line - after all, video cards of this level are always in many ways fashion products that help sell others. So we are waiting for Hawaii.

At one time, AMD engineers were the first to introduce the DVI-D interface, which supports WQHD resolution (2560x1440 pixels). Then DisplayPort and Eyefinity technology, which allows one GPU to display an image on three monitors at once. Now AMD has announced full support for 4K resolution, or Ultra HD. Moreover, the connection to the display can be made using both DisplayPort 1.2 and HDMI 1.4b, but only at a frequency of 30 Hz.

Low-end segment

We talked about the architectural features of Radeon R7/R9 graphics solutions. Now it's time to get acquainted with the video cards themselves. As expected, we divided the devices into three categories: Low-end, Middle-end and High-end. In each segment, there are several video accelerators at once.

Obviously, the creation of graphics processors designed for devices of different classes follows the same scenario. Namely, there is a basic unit - this is the Compute Unit block of the GCN architecture. By adding and subtracting these units, the GPU is obtained.

The second point: for greater clarity and to increase the practical value of this article, we have given examples of specific models graphics adapters, which are in open sale. The main rule is that only non-reference video cards are presented, which, in our opinion, have their own unique features.

And now to business. There are three devices in the Low-end category. Two of them are based on the Oland GPU. The most productive adapter is based on the Cape Verde chip. Specifications Radeon R7 240, Radeon R7 250 and Radeon R7 250X are shown below.