R9 fury nano: AMD Radeon R9 Fury Nano Review

Graphics drivers

Catalyst proprietary graphics driver

Main article: AMD Catalyst

AMD Catalyst is developed for Microsoft Windows and Linux. As of July 2014, other operating systems are not officially supported. This may be different for the AMD FirePro brand based on identical hardware but with OpenGL certified graphics drivers.

AMD Catalyst supports all features advertised for the Radeon brand.

Free open source graphics device driver

radeon

systems. Each driver consists of five parts:

1. DRM Linux Kernel Component
2. Linux Kernel Component KMS Driver: Mainly Device Driver for Display Controller
3. libDRM User Space Component
4. Mesa 3D User Space Component
5. Special and distinct 2D graphics device driver for X.Org Server, which will finally be replaced by Glamor

Free and open source radeon The kernel driver supports most of the features implemented in the Radeon GPU line. ^[6]

The radeon kernel driver is not reverse engineered, but based on the documentation released by AMD. ^[35] These drivers still require proprietary firmware for DRM functionality to work, and some GPUs may not run the X server if one is not available.

Free and Open Source Graphics Device Driver

amdgpu

This new kernel driver is directly supported and developed by AMD. It is available on various Linux distributions and has also been ported to some other operating systems. Only GCN GPUs are supported. ^[6]

AMDGPU-PRO Proprietary Graphics Driver

This new driver from AMD is still under development but can already be used on several supported Linux distributions (AMD officially supports Ubuntu, RHEL/CentOS). ^[36] Driver experimentally ported to ArchLinux. ^[37] and other distributions. AMDGPU-PRO replaces the previous AMD Catalyst driver and is based on free and open source code. 9 «AMDGPU — ArchWiki». wiki.archlinux.org . Received on February 1, 2018.

AMD Radeon R9 Nano

Miniature version of Fury X for compact system units

Content

• Part 1 — Architecture and Architecture 9010 Part 2 — Practical acquaintance
  909 909 901 synthetic benchmarks
• Part 3 — Gaming test results and conclusions 9Nano 4096 MB 4096-bit HBM PCI-E

  Developer Information : ATI Technologies (trademark of ATI) was founded in 1985 in Canada as Array Technology Inc. In the same year it was renamed to ATI Technologies. Headquartered in Markham, Toronto. Since 1987, the company has focused on the release of 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, while AMD GPG remains headquartered at AMD’s former office in Markham, Canada. There is no production. The total number of AMD GPG employees (including regional offices) is about 2,000 people.

  Part 1: Theory and architecture

  Back in June, during the announcement of the Radeon R9 Fury line at E3 2015, AMD told the public that four models of video cards would be made based on the Fiji GPU. The R9 Fury X and R9 Fury have since been released and are on sale, while the other two graphics cards are the R9 Nano models and an as-yet unnamed dual-chip board, about which nothing really is known. But already from the name of the first it is clear that the Nano modification is distinguished by its miniature size, which became possible largely due to the use of HBM standard video memory.

  Well, AMD told even more about the Radeon R9 Nano a few days before today’s full-fledged announcement — the day when these boards are already starting to be sold in retail stores. In August, company representatives revealed to us almost all the theoretical characteristics of the new model, including details about the number of functional blocks. However, AMD decided to postpone the launch of the board, intended for the Mini-ITX systems market, to September 10th.

  If Radeon R9Fury X is designed for the wealthiest 3D gaming enthusiasts who demand maximum performance and efficient water cooling, the Radeon R9 Fury was made for gamers who want traditional air cooling in the form of various original AMD partner coolers. Launching today, the Radeon R9 Nano is purpose-built to provide miniature enthusiasts with the best 3D performance possible in small form factors.

  It seems that life forced AMD to look for interesting solutions and unconventional approaches in order to improve its financial situation. After all, a competitor simply does not have a similar board that is similar in its characteristics (size and power, first of all), so you can try to squeeze more money out of consumers, since interested parties simply have no choice: if you need the most powerful video card that will fit into a small case Mini-ITX, then you have to choose the Radeon R9 Nano, no options.

  At the heart of the Radeon R9 Nano is a full-fledged Fiji GPU based on the GCN 1.2 architecture, which is similar in many details to some of AMD’s previous offerings. Therefore, before reading the theoretical part of the article, it would be desirable to familiarize yourself with our previous materials on video cards of this and past generations:

  • [07/31/15] AMD Radeon R9 Fury: Lightweight replacement for the AMD flagship with HBM support
  • [07/15/15] AMD Radeon R9Fury X: New AMD flagship with HBM support
  • [09/02/14] AMD Radeon R9 285: Tahiti gets 256 bit bus and becomes Tonga
  • [10/28/13] AMD Radeon R9 290X: Reach out to Hawaii! You will receive new vertices of speed and functionality
  • [12.12.11] AMD Radeon HD 7970: new single-processor leader 3D graphics

  “ production

  9HBM, 4096 bits

  to 8.2

  to 8.2

  Theoretical maximum fill rate, gigapixels/s

  Radeon R9 Nano		Code name	9000
  28 nm
  The number of transistors	8. 9 billion
  Architecture	Unified, with the array of general processors for streaming of numerous types of data: vertices, pixels, etc.	Feature Level 12_0 and Shader Model 5.0
  Memory bus	4096-bit: eight memory controllers supporting High Bandwidth Memory
  GPU frequency, MHz	up to 1000 (see below)
  Compute units	integer and float formats, with FP32 and FP64 precision)
  Texture units	256 texture units, with support for trilinear and anisotropic filtering for all texture formats
  ROPs	64 anti-aliasing ROPs with programmable sampling of more than 16 samples per pixel, including FP16 or FP32 framebuffer format. Peak performance up to 64 samples per clock, and in no color mode (Z only) — 256 samples per clock
  Monitor support	Integrated support for up to six monitors connected via DVI, HDMI and DisplayPort
  Memory volume, GB	4
  Memory capacity, GB/s	512
  Calculation productivity (FP32), the Teraflops	to 8. 2
  64.0
  Theoretical texture sampling rate, gigatexels/s	256.0
  The 0007 Radeon R9 Nano takes a completely separate position at the top of AMD’s product line, with an MSRP of $649 — exactly the same as the Fury X and a hundred dollars more than just the Fury. Accordingly, the novelty seems to have rivals in price from the camp of the competing company Nvidia, but they are clearly more powerful and — most importantly! — are not intended for the Mini-ITX market at all, therefore it is very difficult to compare them with each other. If we consider the technical parameters, and not the price, then some of the compact GeForce GTX 9 models come closest in size and performance to AMD’s new product70, like an Asus video card, but they, in turn, already have a significantly lower price with inferior 3D rendering speed. It turns out that the R9 Nano simply has no direct competitors, this model is unique and stands apart on the market. Like other video cards of the elite Fury family, the Nano features 4 GB of video memory, which is enough even for high resolutions at maximum quality settings, although in a number of modern games at 4K resolution and full-screen anti-aliasing is enabled with high quality settings more volume is required. But the first generation of HBM memory simply does not allow making a board with 8 GB of such memory, and nothing can be done about it. However, for a miniature Mini-ITX card, this is not so important, and the closest competitor in the form of GeForce GTX 970 with GDDR5 memory has the same amount of video memory, and with some limitations (in short, it can only access a 3.5 GB partition at full speed). The R9 Nano video card itself is very compact in size, the use of HBM memory made it possible to reduce its physical dimensions and the number of components on the board — in particular, in the power circuit. Since the Radeon R9 Nano video card we are considering today is unique and has many interesting design features, we have included a detailed description of them in a separate chapter. Architectural Features The Radeon R9 Nano graphics card model is based on the full Fiji GPU that we already reviewed in the Radeon R9 Fury X review, and many details can be learned from the related material, including: architectural features, a brief overview of HBM memory , new software features including advanced DirectX 12 support, and more. It is also useful to read the material about the long-known Graphics Core Next (GCN) architecture, which underlies all modern AMD solutions. Although the Fiji GPU has had some changes from last year’s Tonga, it is the third generation of GCN — it can be loosely referred to as version 1.2. The new top-end GPU includes all of the GCN 1.2 enhancements, including improved geometry and tessellation performance, new lossless data compression in the framebuffer, some 16-bit multimedia instructions, and a 2MB larger 2MB cache. level. In terms of computing capabilities, the new GPU received improved scheduling and task distribution and several new instructions for parallel processing of data — all this is described in detail in the Radeon R9 review. Fury X. Schematically, the Fiji GPU is similar to Hawaii, released back in 2013 — both GPUs are divided into four Shader Engines, each of which has its own processor for processing geometric data and a rasterizer, as well as four enlarged ROPs , capable of processing 16 pixels per clock (a total of 64 ROPs in each of these chips). The GPU has a single command processor and eight Asynchronous Compute Engines, which have been modified to reflect the changes in GCN 1.2. Specifically, the Nano model uses a full-fledged version of the Fiji video chip — with all active units, except perhaps with a reduced clock speed of the GPU: , and the total number of stream processors in this model is 4096, like in the older Fury X model. The number of texture units is the same — 256 pieces, as well as the number of geometric blocks and ROP blocks — the last ones in the chip are 64 pieces. All currently released video cards based on the Fiji chip have 4 GB of HBM memory attached directly to the GPU using 4096-bit memory bus. We published a separate material on this memory, in which all its features are considered in as much detail as possible. Since the Fiji video chip in the R9 Nano is exactly the same as in the R9 Fury X, it has improved video data processing units, etc. In particular, the Unified Video Decoder decoding unit can hardware accelerate the decoding of video data in H.265 format ( HEVC) — Fiji chip was the first discrete GPU with such capabilities. Among other advantages of Fiji, we note support for TrueAudio, LiquidVR, Mantle, Eyefinity and FreeSync technologies, which we have repeatedly written about in our articles. In terms of performance, the Nano is quite impressive. Since it uses a full-fledged Fiji chip running up to 1GHz, the R9 Nano can theoretically achieve almost the same performance as the top-of-the-line Fury X, as they both have nearly identical specs including 4096 stream processors, 256 texture units, 64 ROPs and 4096-bit HBM memory. On paper, the only difference between the R9 Fury X and the R9 Nano is the maximum frequency and typical power consumption. If for R9Nano maximum frequency is set at a very high level of 1000 MHz, while for the top-end R9 Fury X, the turbo frequency reaches 1050 MHz with absolutely no problems. Nano seems to select the very best Fiji chips for use in the Nano, which are able to operate at relatively high frequencies with relatively low GPU voltage and power consumption, respectively. At the same time, the functional blocks in Fiji were not cut down absolutely! Which is quite surprising given the R9 Nano’s power system, powered by just one 8-pin connector, and rated for a typical power consumption of just 175W. It seems that, based on the numbers, R9 Nano should almost keep up with Fury X. But in practice, this is only possible if the power consumption of the board does not exceed the limit of 175 W, because Fury X has much higher power — 275 W . Based on the principle of operation of PowerTune, it will be precisely the limit on energy consumption that will limit the performance in the case of Nano. Even specially selected Fiji GPUs cannot run at 1000 MHz in serious tasks with consumption limited to 175 watts — this is too tight a limit for such a powerful GPU. Therefore, the graphics processor of the R9 Nano model in demanding 3D applications will most often operate at a frequency of about 900-920 MHz, which AMD itself confirms. They achieved a typical GPU clock speed in games of 900 MHz, and the actual value will depend on the application and its operating conditions. So, despite the beautiful 1000 MHz figure, the Radeon R9 Nano video card released today is not capable of supporting this frequency, and in reality it will be about 10% slower. Why, then, did the experts set such a high turbo-frequency value? AMD claims that this is done so that the board in any conditions always rests only on the consumption limit, ensuring maximum rendering speed within it. It’s not very clear, however, why the video chip should operate at a frequency of 1000 MHz in undemanding 3D applications, and significantly reduce the frequency where it would be nice to add it. But the figure of 1 GHz is beautiful in any case — especially for such a miniature board with a powerful GPU. Board Design Features The design of the Radeon R9 Nano combines all of AMD’s latest technologies to pack solid power into such a small package. And the new board looks solid, its construction uses quality materials, as in the case of the R9 Fury X. Many parts of the R9 Nano are also made of metal, as is customary in the premium market, and it also has a matte black PCB. And yet, the main distinguishing feature of the Radeon R9Nano is her size. Unsurprisingly, the compact Radeon R9 Nano is perfect for demonstrating one of the benefits of HBM memory. Since four gigabytes of compact chips are installed in packs directly on the GPU core in Fiji, the total area occupied by these microcircuits on a printed circuit board is very small — three times less than previously common solutions. And if even in the top model Radeon R9 Fury X we saw the advantage of the lack of large GDDR5 memory chips on the board, expressed in the relatively modest dimensions of the new flagship, then the R9 modelNano is even smaller — it is designed to install the board in Mini-ITX format systems (6. 7 × 6.7 inches). The overall length of the new PCB is quite impressive — only 6 inches (about 15 cm), which allows the R9 Nano to be used in motherboards and cases of this format. That’s 40% shorter than the previous generation’s top-of-the-line 11-inch Radeon R9 290X: To fit into such a small footprint, AMD moved some of the power system components to the back of the PCB. Additional power for Radeon R9The Nano provides a single 8-pin PCI Express connector, which is used instead of a pair of 6-pin connectors, also due to its comparative compactness. The modest size of the Radeon R9 Nano allows it to be used in small form factor PCs, which are gaining popularity among gamers who want the same power in a small package without losing the ability to play at the highest resolutions at maximum settings. And the 6-inch Radeon R9 Nano is great for that, coming in cases that you can’t fit any other graphics card of similar performance into. Here are just a couple of examples of Small Form Factor (SFF) cases that AMD recommends for use with their new solution: Lian Li PC-Q33 (pictured) and Cooler Master Elite 110. In our opinion, a very interesting option for a home gaming PC for gaming enthusiasts (not to be confused with ultra-enthusiasts and overclockers): The Radeon R9 Nano graphics card is cooled by an air cooler, specially designed to remove 175 watts of heat from such a small board. The only fan is located in the center of a powerful radiator. The design of the cooler is combined: although an open fan is used, it also blows air from the side. Thanks to this solution, about half of the heat from the board is taken out to the outside of the case, although the rest of the heat remains inside — the case cooling system will take care of its output. The radiator in the Nano model consists of two parts, and the main part is made up of the evaporation chamber and heat pipes. The evaporation chamber is made of copper alloy, it removes heat from the Fiji core and the stacks of HBM memory installed on it, and the heat pipes effectively distribute it throughout the radiator area. The second part is a small heatsink with another heatpipe, mounted separately and cooling only the field-effect transistors (MOSFETs) in the voltage regulator module (VRM). The use of the vapor chamber in the R9 Nano is explained by the fact that such solutions are more efficient than simple radiators, in this case, both the vapor chamber and heat pipes are used, which is rarely found together. AMD assures that the cooling system used is very efficient, and the performance of the R9 Nano should not be limited by the GPU temperature limiter. Tiny graphics card is designed to operate at a target GPU temperature of around 75 degrees, 20 degrees below the target temperature set for the previous generation Radeon R9 model290X, and the decrease in frequency and voltage (so-called throttling) will only begin at 85 degrees. In addition, AMD claims that the R9 Nano has a very quiet cooling system — they even compare the noise from it with the natural sound background in libraries. According to the company’s own measurements, the R9 Nano’s efficient air cooler powers today’s top-end GPUs at 42 dBA, a whopping 16 dBA lower than the fairly noisy Radeon R9 290X cooler. By the way, it is quite possible that towards the end of the year, AMD partners will release versions of the Radeon R9Nano, based on boards and coolers of their own design, but at the moment boards are produced exclusively based on the reference version developed by AMD. And we are unlikely to see any unique modifications with a different design on sale before the Christmas holidays. Performance Brief, Positioning and Conclusions Interestingly, AMD itself often compares the new product with the Radeon R9 290X, the previous generation’s top-of-the-line board based on the Hawaii chip, which does not have particularly good energy efficiency and has a loud cooler. It is clear that against such a background R9Nano will look like its complete opposite. For example, in the 3DMark Fire Strike Ultra benchmark and 4K resolution, the R9 Nano delivers noticeably better performance than the Radeon R9 290X and GeForce GTX 970: At the same time, the Radeon R9 290X consumes 30% less energy (175 W versus 250 W), and the cooler of the new product is much quieter than that of the top motherboard of the previous generation. Together this gives R9Nano is a twofold advantage in energy efficiency over the Radeon R9 290X, not to mention the much greater compactness of the new product. Even if we compare the Nano against other models in the Fury line, its performance should be very close, given the approximately 15% lower typical GPU operating frequency compared to the Fury X. That is, the miniature R9 Nano will provide approximately 85% of the speed of a top board in a modern AMD line, and the “airy” Fury on a stripped-down Fiji overtakes it in theory by only 8%. So in terms of speed, the new product should be somewhere between the Radeon R9390X and R9 Fury. However, these models were clearly not AMD’s main target for the release of the R9 Nano. The miniature model is made to be the most powerful Mini-ITX graphics card on the market. Among its compact competitors, the most productive are the Radeon R9 280 and GeForce GTX 970. There is also the GTX 960, but it is much slower. In particular, Asus has a compact version of the GeForce GTX 970 in its lineup, which is closest to the R9 Nano in terms of size and performance. But this Nvidia video card is based on a non-top GM204 chip and is clearly inferior even in theory to a full-fledged Fiji, even if in reality it operates at a frequency of only about 900 MHz. On average, AMD estimates the advantage of their solution at 30% when rendering at 4K resolution (when the GeForce starts to suffer from a lack of video memory): The R9 Nano is significantly faster than both of them. Although the advantage of the Nano will differ in different conditions, on average, the Radeon R9 Nano can be safely called the most powerful among Mini-ITX graphics cards. Interestingly, in the conditions of cramped cases, the R9Nano’s advantage is even higher — due to low heat generation and thoughtful cooling, it does not reduce frequencies when overheated. These slides indicate the loss in orange when going from testing in an open area to a closed tight mini-case. And judging by a couple of gaming tests from AMD, the performance of their new product in small SFF cases practically does not decrease, while the Mini-ITX version of the GTX 970 significantly reduces the frequency of the GPU, which results in the loss of precious frames per second. In general, in terms of computing and gaming performance, everything is fine with the new product, R9 Nano is simply the best in its class, but the price remains an important issue. Especially considering the special selection of the best Fiji video chips, which are quite expensive to manufacture, especially because of the HBM memory. In addition, an advanced cooler with a vapor chamber is installed on the board, which also adds a lot to the cost. And so it’s not surprising that AMD is positioning the R9 Nano as another premium series video card, similar to the Nvidia Titan. After all, R9Nano has unique characteristics, offering hitherto unattainable performance within the Mini-ITX market. Therefore, the price of the new product was set quite high — as much as $649 for the US market, exactly like the older water-cooled Fury X. Let’s hope that this will allow AMD to improve its financial position. True, so far the market for Mini-ITX solutions is rather small compared to the market for large gaming PCs, but there is hope for its further growth. Based on the difficulty of selecting those suitable for work in R9Nano video chips, the question of the availability of the solution on the market also arises — will AMD be able to supply a sufficient number of such boards to retail networks? And here the R9 Nano will help that the market for gaming mini-PCs is not too large, and they should be able to cope with the supply of the video card in question for $650. Moreover, the situation with the production of Fiji chips is constantly improving, more and more of them are being produced, and there is no longer a severe shortage on the market. The Radeon R9 Nano introduces a new class of powerful small-sized graphics cards suitable for installation in mini-ITX chassis. At the same time, the new product is up to 30% faster while consuming 30% less power than the top model of the previous generation Radeon R9290X. Yes, and the top R9 Fury X in terms of speed, it will yield quite a bit. All in all, at just 175W, the Radeon R9 Nano is the most energy efficient and powerful board in the Mini-ITX graphics card market segment. It also has a minimal size and a very efficient air cooler. All together, this allows you to create the most powerful small form factor gaming PCs that were simply not possible before. After getting acquainted with the characteristics and theoretical data of the new video card model Radeon R9Nano, let’s move on to the next part of our article, which is devoted to the study of the rendering speed of a miniature AMD video card in a set of synthetic tests. In continuation of our material, we will compare the performance of AMD’s unique new product both with other motherboards from this company, and with the speed of a pair of Nvidia video cards with similar characteristics. AMD Radeon R9 Nano — Part 2: Card Features and Synthetic Benchmarks → 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000 Nano — $595 (as of 06/01/16) Cons High power consumption Several partner models available Key Features Review Price: £420.00 Core clock 1000 MHz 4 GB High Bandwidth Memory 8.9 billion transistors 3584 stream processors Requires one 8-pin and one 6-pin power connector Manufacturer: AMD What is AMD Radeon R9 Fury? AMD has promised several Fury-named cards at the top of its latest graphics range. The most recent of these is the Radeon R9 Fury, which sits between the flagship Fury X and the tiny Fury Nano. There aren’t many differences between Fury and Fury X when it comes to basic equipment. It’s important to note that the Fury is significantly cheaper than AMD’s current flagship model. AMD Radeon R9 Fury — under the hood The Radeon R9 Fury is not a flagship in the same vein as the Fury X, but it uses pretty much the same hardware. At the heart of it all is the Fiji Core, which uses the familiar Graphics Core Next architecture with turbocharged components and plenty of hardware in just about every department. For example, each shader engine now has 16 stream processors instead of the 11 included in older cards. This offers a huge jump in performance. Roundup: Game of the Year However, the most notable new component is high-speed memory. This replaces GDDR5 on older cards and uses a much wider interface. The high-bandwidth memory inside the Fury has a monstrous 4,096-bit interface that is eight times wider than the 512-bit bus used on older cards. This makes up for the lack of speed in the chips themselves — Fury’s 4 GB of memory runs at 500 MHz, which is slower than any other high-end GPU. There are only a couple of differences between Fury and Fury X. The cheaper card has 3584 stream processors divided into 224 texture units, as opposed to the flagship model’s 4096 stream processors. The Fury runs at 1000MHz, but its big brother runs at 1050MHz. Fury’s main competition is the Nvidia GeForce GTX 980, whose prices start at £384. It has an attractive spec of its own: a 1126MHz core clock that boosts up to 1266MHz, 4GB GDDR5 memory at 7010MHz and 2048 stream processors . AMD Radeon R9 Fury — Score Analysis AMD’s latest card left our 1080p tests in the dust. It ran through most of my test games with average frame rates in excess of 100fps. His best score was in Batman: Arkham Origins where he maxed out at 200fps. Even the lowest frame rate of 82fps in Crysis 3 is still excellent. The Fury 1080p tests also put the Nvidia GeForce GTX 980 ahead of the competition. Radeon beat its competitor in almost every test; it only fell behind in Battlefield 4. Even so, the Radeon result of 85fps was only one frame behind the Nvidia card and is still easily playable. This trend continued at 2,560 x 1440. The gap between AMD and Nvidia was noted in 3DMark, with Fury’s score of 6697 on Extreme easily surpassing the GTX 980’s 5620. 60 frames per second on average. This is the place for perfect playability and five frames faster than the GTX 980. The gap in Batman: Arkham Origins was even bigger. Here the Fury ran at 130fps, which was 25fps faster than the GTX 980 could handle. Fury only crashed in Crysis 3, where its 1440p minimum at 34fps couldn’t match the GTX 980’s 38fps. However, it got the upper hand with an average frame rate of 57fps, which is 11 frames faster than the Nvidia card. Competitive aspects aside, the fury was impressive at 1440p. Every average was perfectly playable, which means there won’t be any stuttering when running the most demanding games. 3840 x 2160 resolution is the current basis for single-screen gaming, and that’s what all high-end cards strive to manage. The AMD card continued to dominate Nvidia. In BioShock Infinite, its 48fps average was seven frames faster than the GTX 980; and in Metro: Last Night, Fury averaged a smooth 41 fps. That’s far beyond 32fps from an Nvidia card. In other 4K games, Fury delivered playable frame rates that the GTX 980 couldn’t. Its average of 29fps in Crysis 3 is almost playable and much more convincing than the 23fps scored by the GTX 980. In Battlefield 4, Fury averages 32fps on the right side of 30fps, and is four frames better than the Nvidia GPU. The latest set of synthetic results highlighted Fury’s position in the market. Its 3DMark: Fire Strike Ultra score of 3639 is well ahead of the GTX 9’s 2 Average current Price (number of proposals) in the Moscow retail: CARE CARKS Competitors R9 Nano — $ 595 (as of 01.06.16) R9 390x — $ 485 (as of 01. 06.16) GTX 980 Ti — $655 (as of 06/01/16) R9 Nano — $595 (as of 06/01/16) R9 Fury X — $788 (as of 01.16) R9 Nano — $595 (as of 06/01/16) , and is even a short distance from the GTX 980 Ti and Fury X, which scored 3888 and 3943 respectively. AMD hardware outperforms the GTX 980 in almost every gaming test, but the Fiji graphics core still lacks performance compared to Nvidia’s Maxwell architecture. The Fury’s idle power consumption of 104W is much less than the GTX 980’s 68W — or even the GTX 980 Ti’s 101W. The Radeon also failed to impress in peak performance. When running at 100%, it consumed 373 watts from the mains. That’s a lot more than the 273W and 330W of the GTX 9.80 and GTX 980 Ti, respectively. Related: Best RPGs AMD Radeon R9 Fury — Things to Consider The high power requirements and relative inefficiency of the Fiji core mean there are fewer versions of Fury for board partners than competing Nvidia hardware. Only four models are currently available. Prices range from £420 to £480, but none of these versions are overclocked — they all run at stock 1000MHz. The most expensive model from XFX, but it’s also the most innovative: it uses a separate water-cooling unit, a design decision borrowed from the faster Fury X. This is in stark contrast to the GTX 980 marketplace. Dozens of models are available for prices ranging from £384 to £500. And here even modest models come with a boost. For example, the £384 MSI card comes with a 1126MHz core boosted to 1152MHz, while the £400 Gigabyte card is overclocked to 1178MHz. The changes made to the GTX 980 won’t be enough to catch the Fury, but they will certainly close the performance gap — many GTX 9 cardsThe 80 comes with free games but is cheaper than the Fury. None of the AMD cards currently ship with gaming kits. Should I buy AMD Radeon R9 Fury? Fury carves its niche in the high quality graphics market. It can comfortably beat the GTX 980 in benchmarks, but it can’t match the GTX 980 Ti and Fury X in terms of performance. This pattern is repeated at the checkout: it’s more expensive than the GTX 980, but significantly cheaper than the other two cards. It’s capable of passing 1080p and 1440p benchmarks, and it’s a reasonable bet at 4K as well — it’s capable of handling more games at playable frame rates than its closest competitor. Overall, this definitely has the biggest hit at £400-450. judgment If AMD and Nvidia flagship cards are too expensive, the AMD Radeon R9 Fury is a good option for high-end gaming. However, be careful — its power consumption is high. Mike has worked as a technology journalist for over a decade and has written for most of the UK’s most prominent websites and magazines. While writing articles about technology, he developed an obsession… Unlike other sites, we thoroughly test every product we review. We use standard industry benchmarks to properly compare features. We will always tell you what we find. We never, ever accept money for a product review. Tell us what you think — send your letters to the editor. News | Page 17 | Computer Technology and Games Nvidia is about to launch one of its fastest GeForce GTX video cards, the video card will have two processors based on two GM 200 chips, such chips are used for the GeForce GTX Titan X and GTX 980 Ti video cards. This news is one of those that were presented at a secret meeting in New York, for the first time, journalists were shown the new two-chip flagship from Nvidia today. Nvidia’s dual graphics cards were the premium class, the company has always strived to provide the fastest video accelerator. As a result, Nvidia had to wait for the fastest and most powerful GM 200 chip before releasing a dual video adapter. Initially, Nvidia released the entire line based on the GM200 and only then prepare for the release of the flagship. The latest GeForce GTX Titan Z had a number of drawbacks, and the main one is the price. Nvidia is preparing a new flagship very high quality so as not to make mistakes and shortcomings when releasing the GeForce GTX Titan Z, the price range was squeezed to the maximum… See full article > NVIDIA has released another version of the GTX 980 graphics card for laptops and it will not differ from the desktop version, the decision to abandon the M-type consoles was deliberate. NVIDIA claims to use the same chip as for PC. However, the GPU chip was selected for the lowest thermal power. The GeForce GTX 980 laptop graphics card will ship with 2048 CUDA cores, 128 TMUs, and 64 ROPs, so it’s a full GM204 chip, no cuts. The most interesting thing is the amount of memory of the video adapter, in the mobile version, so to speak, 8GB will be available. Moreover, NVIDIA claims that the new GTX 980 will run at about the same frequencies as the desktop version of the video card. However, the end frequencies will vary depending on the laptop model, but the base frequency is set to 1064 MHz, while the memory frequency will be 7000 MHz and the bus width will be 256 bits. The bandwidth of the GTX 980 for laptops will be the same as… See full article > Republic of Gamers today announced a new motherboard, Maximus Extreme VIII , a new flagship based on Z170 logic, this motherboard is a shining example from ASUS in creating the gaming motherboard that so many have been waiting for. Maximus Extreme VIII comes with new overclocking technologies and brings you new gaming experience. Based on the new Intel Z170 express chipset and on the new LGA 1151 processor socket, to support the latest 6th generation processors from Intel. Maximus Extreme VIII performs well on the new platform, setting eight world records in the process. Designed and built for the hardcore enthusiast, it comes with a 2.6-inch dashboard for monitoring and extreme overclocking. Like all predecessors of this panel, the novelty will be able to display parameters and provides real-time monitoring of such parameters: processor temperature, coefficient … See full article > The new ASUS ROG Swift PG348Q bvttn Curved Monitor was unveiled at IFA 2015. ASUS has decided to launch a brand new product. The ASUS ROG Swift PG348Q bvttn features an ultra wide 21:9 aspect ratio with 3440 x 1440 resolution. Its 100Hz refresh rate and NVIDIA G-Sync technology will keep your games looking smooth and fluid. There are six preset picture modes so you can customize the visuals to the highest quality for your game genre. The monitor is equipped with DisplayPort, HDMI, four USB 3.0 ports, and powerful built-in stereo speakers. Fully adjustable stand with tilt and swivel, height adjustment and built-in logo light. Specially for ONLY GAMERS and CAPATOB by GRU64rus. See full article > MSI, one of the world’s leading manufacturers of computer components, is proud to announce a revolutionary collaboration with Corsair. This collaboration led to the creation of a gaming graphics card from MSI, with a combined cooling system. The GeForce GTX 980Ti is cooled by a closed liquid cooling loop. This graphics card is an addition to the legendary MSI Hawk family, delivering outstanding gaming performance at revolutionary low temperatures and noise levels. The water block with micro-channels based on copper will take care of efficient heat transfer, assisted by a high-pressure circulation pump. The low-profile aluminum heatsink, which is easy to install, is equipped with a 120 mm Super Silent fan with variable speed depending on the temperature of the GPU. However, to get the best performance, memory and VRM chips need to be equipped with the same… See full article > At IFA 2015, modding champ Ali Abbas built an epic dual-GPU combo-cooled ROG Matrix GTX 980 Ti build upon request. The project is based on the ROG Maximus VIII Extreme motherboard, with an Intel Core i7 6700K processor. Specially for ONLY GAMERS and CAPATOB by GRU64rus. See full article > Razer has become known for delivering high-end gaming peripherals in terms of quality and performance, with a strong focus on the gaming market. Razer has removed the veil of secrecy from the new controller, which has already been nicknamed the Wildcat, designed to win victories in gaming tournaments, as well as for players seeking to compete. This controller joins an already existing line of peripherals designed for Xbox One. Made for tournaments and designed with input from the world’s top esports players, added four additional Multi-Function buttons to select settings, allowing you to change settings on the fly. The controller’s memory allows you to save all possible settings for your games. The cable connecting the set-top box with the controller is placed in a strong sheath and all the grinding is not how much for it. A novelty for such a controller was the appearance from the bottom of additional fully… See full article > The Radeon R9 Nano graphics card is based on a fully unlocked graphics core codenamed Fiji, the GPU has 4096 stream processors, 256 texture units and 64 raster units. The clock speed of the processor core is calculated at 1000 MHz — using the terminology from NVIDIA — in fact, it is not the base frequency, but the Tune Boost frequency, that is, in acceleration, it means the base frequency is slightly lower. Theoretically, this acceleration can reach 1 GHz, but the Tune controller Boost will adjust the frequency both up and down to save R9Nano balance between silence and performance. It remains only to check how AMD Radeon has configured all these compromises. The Tune Boost power values ​​are very aggressive and AMD thinks so, the target temperature is set to 75°, which is not a critical GPU temperature, the critical temperature calculation goes to 90-100 degrees. The temperature of 75° is not chosen… See full article > AMD Radeon R9 graphics card specifications published yesterdayNano, the so-called specification sheet. The published material showed that the card has 4096 stream processors, like the older brother Fury X. The power consumption of the novelty was reduced from 275W to 175W, the frequency of the GPU remains at the bar of 1000 MHz, how did they manage to realize this? Silence about this. AMD Radeon R9 Nano is faster than GTX 970 from pre-test results. Although the GeForce GTX 980 has a TDP of 165W and it will never be in Mini ITX format. One of these options from NVIDIA is only possible based on the GeForce GTX 970. That’s why the GTX 970 is taken as the starting point. But does it make sense? Of course, yes, but it’s worth looking at in comparison with the GTX 980. AMD Radeon R9 Nano can be a worthy answer to the NVIDIA GTX 980/970, but only time has been lost and it will be difficult to master the market. The question of overclocking remains open, the older brother Fury X showed a weak overclocking result… See full article > Everyone who is familiar with modding and even with overclocking should definitely hear about the modder by the nickname 8Pack, who is one of the best extreme overclockers in the world, and currently holds numerous world records. 8Pack represents the overclockers in the UK, so you can find him on the overclocking forums offering his expertise to fellow overclockers. Let’s take a look at his latest creation. Characteristics of this unique assembly. — Processor: i7-5960K overclocked to 4.625GHz — Motherboard: ASUS RAMPAGE V EXTREME -Memory: 32GB 2666MHz Kingston HyperX Predator DDR4 — Video cards: 4 x ASUS GTX Titan X 12GB — Hard drives: SSD 2 x 1 TB Samsung 850 Pro in RAID 0 configuration Case: Lian Li PC-D 600 series -Cooling: Water-cooling components from the Slovenian company EK are used — Power supply: SuperFlower 2000W PSU, in theory — with such a powerful graphics system, you need to use two power supplies, so… View full article > I bought a gt720 video card to replace my gts250, I thought it would be more powerful, in the end it turned out to be much weaker, what should I do? See full article > AMD’s new Fiji graphics card family is one of the most hotly anticipated graphics processors, making an innovative push for high-end new generation technology to form a new direction for the gaming industry. There is an unspoken rule among engineers that the percentage of new parts in a product as a whole should not be 20-30 percent, but AMD decided on fifty percent. They completely changed the entire memory architecture and the bus. The new memory was supposed to provide a higher level of bandwidth than the memory with traditional technology. The Fury is the most interesting product, being just a slightly stripped-down version of the full Fiji core, not much less powerful and air-cooled. The Fury core has 3,584 stream processors running at 1000 MHz and 224 texture units, while Fiji has 4096 stream processors running at 1050 MHz and 256 texture units. This is where the differences end, the memory of the two video cards … See full article > In WIN is a well-known company for the production of cases for various computer systems. Stylish mid-tower cases, 805 series for enthusiasts are launched. The new case is built from high quality materials and provides rich options for expanding, cooling and configuring your builds. A new stylish series of cases from In Win 805 Mid-Tower will be made of aluminum and tempered glass 3 mm thick. This combination is unique for the company. The front panel will have a distinctive feature, this highlight, according to the plan of In WIN, should be the ability to see through the frame slots in the form of honeycombs the internal structure and filling of the case, and the glass sides are not far behind in the stylish look of the entire filling. The panel also houses one USB 3.0 port, two USB 2.0 connectors, as well as audio input and output. You’ll have to sweat to fix everything and lay it down. The case will be able to provide excellent air circulation and cooling of modern… See full article > Today NVIDIA launches a new graphics card GeForce GTX 950 based on the Maxwell core. The GeForce GTX 950 graphics card is aimed at those users who play on graphics cards with low cost, the price is as low as $200 and with a monitor resolution of no higher than 1080p. The GeForce GTX 950 is the second generation of the Maxwell architecture in the sub-$200 price range. The first generation of Maxwell were the GTX 750 and GeForce GTX 750 Ti graphics cards. NVIDIA decided to update its lineup of video cards and released the GTX 950 brings improvements in terms of performance, power efficiency and gives users more options, which include full support for the DirectX 12 API. NVIDIA GeForce GTX 950 Reference PCB: The NVIDIA GeForce GTX 950 GPU is based on the GM206-250-A1 graphics core, which is a stripped-down version of the GM206 core. The GTX 950 graphics card has 768 CUDA cores, 48 ​​texture units, 32 raster units. Reference video card… See full article > The first results of testing a new video card from NVIDIA, this is the GeForce GTX 950, got into the network, published the results on a Chinese resource. It has become a habit that all new products are tried by the Chinese and only then by the whole world. Let’s take a look at what’s new. For experiments, Gainward GeForce GTX 950 designed for the Chinese market was presented. In the 3DMark test, the frequency of the video card fluctuated around 1430 MHz, this frequency is due to a non-reference version of the adapter, while the result of the video adapter does not cause shock. As expected GTX 950 is not ahead of 960 but higher than 760. The lead is not large and is within the margin of error, so we’ll wait for the video card and see the performance ourselves. In the photo, the Gainward GeForce GTX 950 video card with a back clip is rumored to be preparing eight video adapter models from different manufacturers, even with 4 GB of on-board memory. Specially for ONLY GAMERS and CAPATOB by GRU64rus. See full article > Nvidia is releasing a new king early this year, every enthusiast’s dream, perfect for 4K monitors, GTX Titan X with 3072 CUDA cores priced at a thousand US dollars. But a few months later, the throne under TitanX was shaken with the release of the new GTX 9 graphics card.80 Ti priced at $650 which is perfect for gaming. So the time for reference solutions is over and many Nvidia partners have released their interesting solutions, quieter and colder. Testing of video cards has passed and there is a lull, let’s take a look at the rival from AMDRadeon R9 Fury X. Let’s put two GTX 980 Ti in 2way SLI mode and two Radeon R9 Fury X in AMD CrossFireX mode. We will test performance in games, in heavy 4K mode. Although each of the rivals can provide fps no higher than 30 fps with game settings no higher than average, let’s look at the performance of a pair of video adapters, because sli and CrossFireX in 4K should best provide maximum fps. Testing will be in two modes, at nominal frequencies and at … See full article > We have more detailed information about the new GeForce GTX 950, GTX 950Ti and some original photos from one reliable resource. It became known about another video adapter from NVIDIA, to fill a niche between the 960 and 950 video cards. It is safe to confirm the exact specifications of the NVIDIA GeForce GTX 950 graphics card from official documents, GTX 950 has 768 CUDA cores and 2 GB of GDDR5 memory with 128-bit memory bus. GeForce GTX 950 is pre-shipped with a base frequency of 1089 MHz and a frequency at Boost 2.0 of 1266 MHz. Interestingly, the memory frequency will not be 7 GHz, but slightly below 6.6 GHz (1652 MHz). This is not much less than the 960 at 100 MHz. This is not a problem for overclockers, you can always overclock. The card will have additional power through one 6-pin power connector and the thermal power of the video card will be 90 watts. NVIDIA is working on another graphics card based on the GM206 processor that will fill the market between GeForce GTX 950… See full article > EK launches a new water cooling product, the applied design solutions will be universal for many cooling components. EK from Slovenia is updating its product line. EK-Predator translates as a predator, this is an unattended water cooling system for intel processors, support for AMD processors is promised but did not specify when. The EK-Predator is pre-assembled and filled with liquid and comes in two versions; 240mm 360mm and provide efficient CPU cooling. EK-Predator uses standard G1/4 threaded holes and 10/16mm Zero-Maintenance Rubber Tube (EK-ZMT), all parts are market supplied and in case of any breakage or expansion, if you decide to add a video card water block to the circuit, all can buy. The EK-Predator kit features an integrated quick connector that allows you to quickly and easily drain fluid or add cooling components. Main Specifications… See full article > A third-party manufacturer or even an ally of the TUL concern, which occupies a leading position in the production of AMD video adapters and operates under the PowerColor trademark, has launched the production of the Devil Radeon R9 390X video card. The video accelerator is equipped with eight gigabytes of on-board GDDR5 memory with a 512-bit bus, as well as 2816 stream processors. The Radeon R9 390X Devil video card is slightly overclocked compared to its reference counterpart. The processor core has a frequency of 1100 MHz, the memory frequency is 1525 MHz, the printed circuit board has been completely redone, the power system looks like eight phases built according to the 6 + 1 + 1 scheme. The novelty should support all modern graphics standards, namely DirectX 12 and such technologies from AMD: Virtual Super Resolution, FreeSync, Liquid VR, 4K. The highlight of this series was the hybrid cooling system Radeon R9390X Devil, this should have a positive effect on the cooling of the new product and on the noise emitted by the video adapter. Specially for ONLY GAMERS and CAPATOB by GRU64rus. See full article > Back in June 2015, AMD launched the production of Radeon P9 FuruX graphics cards, which showed its redesigned Fiji XT chip. Since then, many have tried to overclock this product less or more successfully. Let’s look at one such example of conversion for overclocking with liquid nitrogen. Preparing for interventions. Here are some tips for overclocking the Radeon P9FuruX 1. Use the HDMI port connector (I used it with an HDMI — DVI adapter) on my card, the temperature was -180*C. With a DVI connector, the screen flickers and the card failed to start. 2. If you cannot enter the OS at low temperature, increase the voltage from 0.95V to 1.25V to 1.2V. 3. When the voltage is 1.25V, the temperature should be within -120 -160*s range. 4. When the temperature drops about fifty degrees, keep about 1.35-1.5V (stress) to boot into the OS, then it will be about 1.1-1.15 in standby (0.9V in standby mode). 5. You can install on… See full article > Modding beautiful water-cooled projects, everything is color-matched from computer components to liquid. Here is the list of components: Processor — Core i7 4790K Motherboard — Asus Sabertooth Z97 Mark C Video Cards — EVGA 980 Ti SC + ACX 2.0+ SLI Memory — HyperX 1866MHz Power Supply — EVGA SuperNova G2 1300W Enclosure — Phanteks Enthoo Primo SSD drive — Samsung 850 EVO 1TB x2 + 500GB Listing some parts of the cooling system: Radiators/Fans: ALPHACOOL Monsta 360mm with 6 EK Vardar F2-120 1450rpm, EK-Intensive Cooling CE 420 with 3 Fractal Design Venturi HP 14 and 3st Noctua NF-A14 EK-Cool XE Flow 240 with 2 Fractal Design Venturi HP 12 and 2x Noctua NF-F12. pump reservoir: EK-D5 XRES PWM + EK-RES X3 — TUBE 250 GPU waterblocks: EK-FC Titan X — nickel. CPU waterblock: EK-Supremacy EVO White EditionEK, E.K. PrimoFlex hose 12 / 16mm Specially for ONLY GAMERS and CAPATOB by GRU64rus. See full article > Radeon Rx 300 — frwiki.wiki The Radeon RX 300 series is a series of video cards manufactured by AMD that replaced the RX 200. They correspond to the code name Caraibean Islands and are designed to compete with Nvidia with its GeForce G900 Maxwell. Execution units Range Radeon Templates Radeon R7 360 Radeon R7 370 Radeon R9 380 Radeon R9 380X Radeon R9 390 Radeon R9 390X Radeon R9 Nano Radeon R9 Fury Radeon R9 Fury X Radeon PRO Duo Fineness of engraving 28 nm Chip code Tobago Trinidad Antigua Grenada Fiji Capsaicin Architecture GCN2 GCN GCN3 GCN2 GCN3 Chip surface 160 mm² 212 mm² 366 mm² 438 mm² 596 mm² 2x 596 mm² Number of transistors 2. 08 g 2.8 g 5G 6.2 g 8.9 g 8.9 g Frequency 1050 MHz 925 MHz 970 MHz 970 MHz 1000 MHz 1050 MHz 1000 MHz 1000 MHz 1050 MHz 1000 MHz Turbo frequency — 975 MHz — — — — — — — — Shaders 768 1024 1792 2048 2560 2816 4096 3584 4096 2x 4096 Texture blocks 48 64 112 128 160 176 256 224 256 2x 256 ROP 16 32 32 32 64 64 64 64 64 2×64 L2 cache 512 KB 512 KB 512 KB 512 KB 1024 KB 1024 KB 2048 KB 2048 KB 2048 KB 2048 KB Memory type GDDR5 GDDR5 GDDR5 GDDR5 GDDR5 GDDR5 HBM HBM HBM HBM Memory capacity 2 GB 4 GB / 2 GB 4 GB / 2 GB 4GB 8 GB 8 GB 4GB 4GB 4GB 2x 4 GB Memory bus width 128 bit 256 bit 256 bit 256 bit 512 bit 512 bit 4096 bits 4096 bits 4096 bits 2x 4096 bits Memory bandwidth 96 GB/s 179 GB/s 176 GB/s 182. 4 GB/s 384 GB/s 384 GB/s 512 GB/s 512 GB/s 512 GB/s 2x 512Gbps Memory frequency 1500 MHz 1400 MHz 1375 MHz 1425 MHz 1500 MHz 1500 MHz 500 MHz 500 MHz 500 MHz 500 MHz Thermal envelope (TDP) 100 W 110 W 190 W 190 W 275 W 275 W 175 W 275 W 275 W 350 W Calculation FP16 — — 3.4 teraflops 3.9 teraflops — — 8.2 teraflops 7.1 teraflops 8.6 teraflops 16.4 teraflops Calculation FP32 1.6 teraflops 1.9 teraflops 3.