KitGuru KitGuru.net – Tech News | Hardware News | Hardware Reviews | IOS | Mobile | Gaming | Graphics Cards
Today we take a somewhat belated look at the latest AMD flagship, the R9 Fury X. This graphics card has been designed to tackle Nvidia in the high end, specifically their similarly priced GTX980 Ti. The Fury X is the first consumer GPU equipped with Stacked High Bandwidth Memory (HBM) – but the big question needs answered … how does it fare against the latest Nvidia solutions?
First let us look at how the Fiji based Fury X slots into the current AMD line up.
| GPU | R9 390X | R9 290X | R9 390 | R9 290 | R9 380 | R9 285 | Fury X |
| Launch | June 2015 | Oct 2013 | June 2015 | Nov 2013 | June 2015 | Sep 2014 | June 2015 |
| DX Support | 12 | 12 | 12 | 12 | 12 | 12 | 12 |
| Process (nm) | 28 | 28 | 28 | 28 | 28 | 28 | 28 |
| Processors | 2816 | 2816 | 2560 | 2560 | 1792 | 1792 | 4096 |
| Texture Units | 176 | 176 | 160 | 160 | 112 | 112 | 256 |
| ROP's | 64 | 64 | 64 | 64 | 32 | 32 | 64 |
| Boost CPU Clock | 1050 | 1000 | 1000 | 947 | 970 | 918 | 1050 |
| Peak GFLOPS (SP) | 5914 | 5632 | 5120 | 4849 | 3476 | 3290 | 8602 |
| Memory Clock | 6000 | 5000 | 6000 | 5000 | 5700 | 5500 | 500 |
| Memory Bus (bits) | 512 | 512 | 512 | 512 | 256 | 256 | 4096 |
| Max Bandwidth (GB/s) | 384 | 320 | 384 | 320 | 182.4 | 176 | 512 |
| Memory Size (MB) | 8192 | 4096 | 8192 | 4096 | 4096 | 2048 | 4096 |
| Transistors (mn) | 6200 | 6200 | 6200 | 6200 | 5000 | 5000 | 8900 |
| TDP (watts) | 275 | 290 | 275 | 275 | 190 | 190 | 275 |
The Fury X is built on the full Fiji die, on the 28nm process. The core architecture is based on GCN 1.2 which is found in the Tonga/Antigua based R9 285 and R9 380. The Fury X is a scaled up version of this architecture with more texture units and shading cores. Compute is a primary focus for AMD with their Fiji architecture – while for instance a card such as the R9 390X is equipped with 11 Compute Units, the Fury X has a whopping 16.
Fiji has four Shader Engine Clusters, like the older Hawaii GPU but this time it is based around 16 CGN Compute Units, as we said – an increase from the 11 per cluster in Hawaii. Fiji has 4,096 stream processors, 256 TMU's and 64 ROPS. Double precision support is slower than the 390x – running at a mere 1/16th of SP.
Analysing the hardware proves interesting, because it is clear that the design is focused on shading pixels and is clearly more powerful than current Nvidia architecture. Nvidia have advantages in other areas however such as final output blending and situations which would call for heavy geometry loads.
AMD have been keen to address power consumption. According to engineers related to the project, the Fiji design utilises improved SV12 power gating. Incorporating a liquid cooler onto the Fury X also means that temperatures will be reduced, having the knock on effect of reducing power consumption. Reduced operating temperature allows the GPU to lose less power, hence reducing the overall power consumption level.
The adoption of low latency HBM memory will have a positive impact as it has a reduced power demand over GDDR5 memory. This memory is running at a super high bandwidth through a slower bus, connected to the GPU via an interposer. It also reduces the physical footprint required for a PCB design meaning smaller graphics cards now, and in the future. HBM memory forces data through an insanely wide 4096 bit bus while the memory is clocked at a very modest 500mhz. This configuration allows for incredible memory bandwidth performance, rated at 512GB per second.
As we have detailed in previous articles, first generation HBM is limited to four 1GB stacks for a total of 4GB. This could immediately be seen as a possible concern, especially as Nvidia's GTX980 Ti and Titan X are equipped with 6GB and 12GB of GDDR5 respectively. Not only that, but the lower cost AMD 390 (review HERE) and 390X (review HERE) are equipped with 8GB – or double the count of the more expensive, flagship Fury X flagship card.
We know that many of the latest game engines such as Grand Theft Auto 5 will reap reward from larger framebuffers at Ultra HD 4K resolutions, especially when the image quality is cranked. Generally, more memory is a good thing to see, although our tests have shown that a card such as the R9 390 won't really see much benefit from 8GB of GDDR5 at 4K – the core itself runs out of horsepower before the full 8GB can be fully utilised.
While the somewhat modest allocation of 4GB of HBM on Fury X could be seen as a major concern, AMD have stated that the buffer on this card is used as efficiently as possible, and AMD's Robert Hallock stated that 4GB is more than enough for games right now, and in the future. That said, you can be sure the second generation of HBM will likely double capacity on upcoming flagship graphics cards.
So what do I have in store for you today? For the last 9 days or so I have been running a series of tests at both 1440p and Ultra HD 4K resolutions – with the latest AMD and NVIDIA drivers – to keep everything on a completely even footing. It is time consuming, but worth it. All AMD video cards are tested with the 15.6 Catalyst Beta, and all Nvidia cards the Forceware 353.30 driver.
Our Visiontek R9 Fury X 4GB sample arrived with us just after the official AMD launch a short while ago. We decided to take more time and to retest all hardware with the latest AMD and Nvidia drivers – rather than rushing to get a review up as soon as possible with some comparison cards running on older driver revisions. Special thanks to Visiontek and (many) other AMD partners who offered us samples for this review.
The R9 Fury X sample is protected inside thick slabs of foam. This is a self enclosed liquid cooled unit, so additional protection is needed to prevent leaks due to rough handling during shipping.
Not an extensive bundle, but the DVI converter cable will likely prove handy for many people using an older screen without HDMI or DisplayPort connectors.
The Fury X 4GB card is very tiny, thanks to the HBM implementation. It measures around 190 millimeters long, although additional mounting space will be needed in the case to fit the 120mm radiator and fan. The R9 Nano – a future release, is said to be only 155mm long!
The radiator is compact and won't be difficult to install close to the card – although we can see this causing some issues for people. Many of our readers have a 120mm or 140mm ‘all in one' CPU cooler already fitted in the rear exhaust position, so this would need to moved out of the way for the R9 Fury X radiator to be fitted.
The R9 Fury X is nicely designed and we like the surface material, which feels great in the hand. It is a multi piece aluminum die cast construction with ‘black soft touch texture side plates'. The Black Nickel aluminum contrasts well with the other materials. There is talk of partners using their own custom faceplates in future.
The card takes power from two 8 pin PCI-e connectors. Above these two connectors is a row of LED lights which are either red or green in operation. A single green LED indicates that AMD ZeroCore is active. The more red LED's are showing, the more active the GPU is. The ‘RADEON' logo on the side also lights up in RED when the system is powered up.
AMD have ditched VGA and DVI connectivity on the rear I/O plate – opting for DisplayPort 1.2a and HDMI ports. We have no problems with this, especially as the card ships with a DVI converter cable, however there is one, rather niggling issue. AMD are still using HDMI 1.4a ports on their cards which are limited to 30hz at Ultra HD 4K.
If you want to game on a large Ultra HD 4K television it is likely it won't have a DisplayPort connector – only a small percentage do. You are therefore not able to get 60hz at the native resolution … and no one wants to play games at 30hz. Nvidia have had full HDMI 2.0 support now for quite some time. This may not be an issue for many, but we have already noticed enthusiast gamers complaining about AMD's lack of HDMI 2.0 support, on our Facebook page. I think this is a rather glaring oversight by AMD.
Getting access to the insides is rather easy and we can see the cooling system is created for AMD by Cooler Master. An all in one pump is in direct contact with the copper plate which covers the GPU core and the HBM memory underneath. Additionally there is a copper pipe and heatsink which covers the VRM components on the PCB. There is a dual BIOS switch, so it is easy enough to flash BIOS 2 and fall back to BIOS 1 in case of problems.
An overview of the Fury X in GPUz – as discussed in detail on the previous page of the review.
As we spent a lot of time retesting all our cards with the latest AMD and NVIDIA drivers we felt it was worthwhile adding in results from the Fury X 4GB with it overclocked to breaking point. How far did we manage to get it clocked?
Right now overclocking the memory proves difficult – the option is locked by default. As our friends over at Hardware.info spotted – there are times when the slider for the memory settings seemed to unlock on random boot up. Overclocking the memory caused BSOD's on our particular review system so I left it alone and perhaps will revisit at a later date. For now however, we managed to overclock the core by 80mhz to 1,130mhz. Increasing the power limit past 35% didn't help improve the GPU clocks past 1,130mhz.
All our graphs today will show performance from the Fury X 4GB at ‘out of the box’ speeds, along with a yellow arrow pointing upwards to indicate additional results at 1130mhz.
On this page we present some high resolution images of the product taken with a Canon 1DX and Canon 28-70mm F2.8 lens. These will take much longer to open due to the dimensions, especially on slower connections. If you use these pictures on another site or publication, please credit Kitguru.net as the owner/source.
For the last 9 or 10 days we have been testing and retesting all the video cards in this review with the latest 15.6 Catalyst and 353.30 Forceware drivers. We have also selected some new game sections to benchmark during our ‘real world runs’.
If you want to read more about our test system, or are interested in buying the same Kitguru Test Rig, check out our article with links on this page. We are using an Asus PB287Q 4k and Apple 30 inch Cinema HD monitor for this review today.
Due to reader feedback we have changed the 1600p tests to 1440p, and we have also disabled Nvidia specific features such as Hairworks in The Witcher 3: Wild Hunt as it can have such a negative impact on partnering hardware.
Anti Aliasing is also now disabled in our tests at Ultra HD 4K as readers have indicated they don’t need it at such a high resolution.
If you have other suggestions please email me directly at zardon(at)kitguru.net.
Cards on test:
Visiontek Radeon R9 Fury X 4GB (1,050mhz core / 500mhz memory) & (1,130mhz core)
Sapphire R9 295X2 (1,018 mhz core / 1,250 mhz memory)
Nvidia Titan Z (706mhz core / 1,753 mhz memory)
Gigabyte GTX980 Ti G1 Gaming (1,152mhz / 1,753 mhz memory)
Nvidia Titan X (1,002 mhz core / 1,753 mhz memory)
Nvidia GTX 980 Ti (1,000 mhz core / 1,753 mhz memory)
Asus GTX980 Strix (1,178 mhz core / 1,753 mhz memory)
Sapphire R9 390 X 8GB (1,055 mhz core / 1,500 mhz memory) & (1,144mhz core / 1631 mhz memory)
Sapphire R9 390 Nitro 8GB (1,010 mhz core / 1,500 mhz memory) & (1,125mhz core / 1637 mhz memory)
Sapphire R9 290 X 8GB (1,020 mhz core / 1,375 mhz memory)
Asus R9 290 Direct CU II ( 1,000 mhz core / 1,250 mhz memory)
Asus R9 285 Strix (954 mhz core / 1,375 mhz memory)
Palit GTX970 (1,051 mhz core / 1,753 mhz memory)
Software:
Windows 7 Enterprise 64 bit
Unigine Heaven Benchmark
Unigine Valley Benchmark
3DMark Vantage
3DMark 11
3DMark
Fraps Professional
Steam Client
FurMark
Games:
Grid AutoSport
Tomb Raider
Grand Theft Auto 5
Witcher 3: The Wild Hunt
Metro Last Light Redux
We perform under real world conditions, meaning KitGuru tests games across five closely matched runs and then average out the results to get an accurate median figure. If we use scripted benchmarks, they are mentioned on the relevant page.
Game descriptions edited with courtesy from Wikipedia.
Futuremark released 3DMark Vantage, on April 28, 2008. It is a benchmark based upon DirectX 10, and therefore will only run under Windows Vista (Service Pack 1 is stated as a requirement) and Windows 7. This is the first edition where the feature-restricted, free of charge version could not be used any number of times. 1280×1024 resolution was used with performance settings.
This older Direct X 10 benchmark places the R9 Fury X ahead of the manually overclocked Sapphire R9 390X Tri-X 8GB, and behind the Nvidia GTX980 Ti. When manually overclocked to the limit, the R9 Fury X score increases by around 1,000 points.
3DMark 11 is designed for testing DirectX 11 hardware running on Windows 7 and Windows Vista the benchmark includes six all new benchmark tests that make extensive use of all the new features in DirectX 11 including tessellation, compute shaders and multi-threading. After running the tests 3DMark gives your system a score with larger numbers indicating better performance. Trusted by gamers worldwide to give accurate and unbiased results, 3DMark 11 is the best way to test DirectX 11 under game-like loads. If you want to learn more about this benchmark, or to buy it yourself, head over to this page.
The graphics score of 19,176 points is further increased to 20,358 points after manually overclocking the core to 1,130mhz. Even with the manual overclock, the score falls around 2,000 points behind the reference clocked GTX980 Ti.
3DMark is an essential tool used by millions of gamers, hundreds of hardware review sites and many of the world’s leading manufacturers to measure PC gaming performance.
Futuremark say “Use it to test your PC’s limits and measure the impact of overclocking and tweaking your system. Search our massive results database and see how your PC compares or just admire the graphics and wonder why all PC games don’t look this good.
To get more out of your PC, put 3DMark in your PC.”
The Fury X 4GB scores well in the latest Direct X 11 benchmark – 3,850 points- placing it just behind the reference GTX980 Ti. When overclocked to the limit the score increases to 4,069 points, just ahead of the reference clocked GTX980 Ti and almost level with the reference clocked Titan X. The Gigabyte GTX980 Ti G1 gaming on the other hand scores 4,401 points, putting it way ahead of the manually tweaked R9 Fury X.Unigine provides an interesting way to test hardware. It can be easily adapted to various projects due to its elaborated software design and flexible toolset. A lot of their customers claim that they have never seen such extremely-effective code, which is so easy to understand.
Heaven Benchmark is a DirectX 11 GPU benchmark based on advanced Unigine engine from Unigine Corp. It reveals the enchanting magic of floating islands with a tiny village hidden in the cloudy skies. Interactive mode provides emerging experience of exploring the intricate world of steampunk. Efficient and well-architected framework makes Unigine highly scalable:
- Multiple API (DirectX 9 / DirectX 10 / DirectX 11 / OpenGL) render
- Cross-platform: MS Windows (XP, Vista, Windows 7) / Linux
- Full support of 32bit and 64bit systems
- Multicore CPU support
- Little / big endian support (ready for game consoles)
- Powerful C++ API
- Comprehensive performance profiling system
- Flexible XML-based data structures
We test at 2560×1440 with quality setting at ULTRA, Tessellation at NORMAL, and Anti-Aliasing at x2.
Tessellation performance has improved for AMD with this generation of hardware, although they are still behind Nvidia. The reference clocked GTX980Ti is still around 11 frames per second ahead (average) of the Fury X at 1,130mhz.
Grid Autosport (styled as GRID Autosport) is a racing video game by Codemasters and is the sequel to 2008′s Race Driver: Grid and 2013′s Grid 2. The game was released for Microsoft Windows, PlayStation 3 and Xbox 360 on June 24, 2014. (Wikipedia).
We test with the image quality on ULTRA and 8x anti aliasing enabled.
At 1440p, The Fury X struggles to even match the GTX970 in Grid Autosport. This is a poor showing from the AMD flagship, but we will retest at Ultra HD 4K on the next page.
Grid Autosport (styled as GRID Autosport) is a racing video game by Codemasters and is the sequel to 2008′s Race Driver: Grid and 2013′s Grid 2. The game was released for Microsoft Windows, PlayStation 3 and Xbox 360 on June 24, 2014. (Wikipedia).
We test with the image quality on ULTRA and 8 anti aliasing enabled.
At Ultra HD 4K, the Fury X scores much better against the competition. It is still around 7 frames per second behind the GTX980Ti, but in real world terms, the differences in frame rate are barely noticeable. It is worth pointing out that the Fury X minimum frame rate is almost the same as the reference GTX980 Ti.
The massively overclocked Gigabyte GTX980 Ti G1 Gaming is leagues ahead of any other single GPU card.
Tomb Raider received much acclaim from critics, who praised the graphics, the gameplay and Camilla Luddington’s performance as Lara with many critics agreeing that the game is a solid and much needed reboot of the franchise. Much criticism went to the addition of the multiplayer which many felt was unnecessary. Tomb Raider went on to sell one million copies in forty-eight hours of its release, and has sold 3.4 million copies worldwide so far. (Wikipedia).
We test at 1440p with the ‘ULTIMATE’ image quality profile selected.
Tomb Raider is perfectly playable at 1440p, averaging between 80 and 90 frames per second and holding a minimum frame rate above 60 at all times. Again, the reference clocked GTX980Ti scores a little better.Tomb Raider received much acclaim from critics, who praised the graphics, the gameplay and Camilla Luddington’s performance as Lara with many critics agreeing that the game is a solid and much needed reboot of the franchise. Much criticism went to the addition of the multiplayer which many felt was unnecessary. Tomb Raider went on to sell one million copies in forty-eight hours of its release, and has sold 3.4 million copies worldwide so far. (Wikipedia).
We test at 3840×2160 (4K) with the ‘ULTIMATE’ image profile selected. We normally reduce the image quality profile to ‘ULTRA’ at this resolution, but we decided to keep it at the highest image quality possible.
At Ultra HD 4K, the differential between the reference GTX980 Ti and Fury X are almost completely negated. The Fury X responds well to overclocking, pushing the final performance figures to Titan X levels.
The Witcher 3: Wild Hunt (Polish: Wiedźmin 3: Dziki Gon) is an action role-playing video game set in an open world environment, developed by Polish video game developer CD Projekt RED. The Witcher 3: Wild Hunt concludes the story of the witcher Geralt of Rivia, the series’ protagonist, whose story to date has been covered in the previous versions. Continuing from The Witcher 2, the ones who sought to use Geralt are now gone. Geralt seeks to move on with his own life, embarking on a new and personal mission whilst the world order itself is coming to a change.
Geralt’s new mission comes in dark times as the mysterious and otherworldly army known as the Wild Hunt invades the Northern Kingdoms, leaving only blood soaked earth and fiery ruin in its wake; and it seems the Witcher is the key to stopping their cataclysmic rampage. (Wikipedia).
We test with the highest image quality settings, although I have disabled the Nvidia Hairworks option specifically as it does kill frame rate on many cards. Graphics Preset is on ULTRA and Postprocessing is on HIGH.
I have played The Witcher 3 for around 85 hours and I have completed the single player campaign. I tested the game today by playing 4 different save game stages for 5 minutes each, then averaging the frame rate results for a real world indication of performance – one of the map sections we tested is one of the most demanding in the game and our results can be considered strictly ‘worst case'. The Witcher 3 is a dynamic world, so it is important to run tests multiple times to remove any discrepancies. Our results below are
This is one of the greatest PC games ever released in my opinion, so I spent around a total of 48 hours benchmarking it for this review alone – it should be on your must have list, if you don't have it already.
Performance at 1440p is quite good, if a little slower than the reference GTX980 Ti. Still, the game is perfectly playable and smooth throughout. A simple drop of one or two image quality settings would improve frame rates further.
Notice the R9 Fury X minimum frame rates of more than 40 FPS – a number which other single-GPU AMD cards cannot reach. That's very important for users of an AMD FreeSync 1440P gaming monitor, many of which have a lower FreeSync boundary of 35 or 40 FPS.
The Witcher 3: Wild Hunt (Polish: Wiedźmin 3: Dziki Gon) is an action role-playing video game set in an open world environment, developed by Polish video game developer CD Projekt RED. The Witcher 3: Wild Hunt concludes the story of the witcher Geralt of Rivia, the series’ protagonist, whose story to date has been covered in the previous versions. Continuing from The Witcher 2, the ones who sought to use Geralt are now gone. Geralt seeks to move on with his own life, embarking on a new and personal mission whilst the world order itself is coming to a change.
Geralt’s new mission comes in dark times as the mysterious and otherworldly army known as the Wild Hunt invades the Northern Kingdoms, leaving only blood soaked earth and fiery ruin in its wake; and it seems the Witcher is the key to stopping their cataclysmic rampage. (Wikipedia).
We test with the highest image quality settings, although I have disabled the Nvidia Hairworks option specifically as it does kill frame rate on many cards. Graphics Preset is on ULTRA and Postprocessing is on HIGH.
I have played The Witcher 3 for around 85 hours and I have completed the single player campaign. I tested the game today by playing 4 different save game stages for 5 minutes each, then averaging the frame rate results for a real world indication of performance – one of the map sections we tested is one of the most demanding in the game and our results can be considered strictly ‘worst case'. The Witcher 3 is a dynamic world, so it is important to run tests multiple times to remove any discrepancies. Our results below are
This is one of the greatest PC games ever released in my opinion, so I spent around a total of 48 hours benchmarking it for this review alone – it should be on your must have list, if you don't have it already.
Ultra HD 4K has proven too much of a challenge for the hardware on test, at these image quality settings. The Fury X scores extremely well at Ultra HD 4K however, managing to slightly outperform the Titan X even by 1-2 frames per second. Reductions in image quality would really be needed however to get the frame rates consistently above 30 at all times.Grand Theft Auto V is an action-adventure game played from either a first-person or third-person view. Players complete missions—linear scenarios with set objectives—to progress through the story.
Outside of missions, players can freely roam the open world. Composed of the San Andreas open countryside area and the fictional city of Los Santos, the world of Grand Theft Auto V is much larger in area than earlier entries in the series.
The world may be fully explored from the beginning of the game without restrictions, although story progress unlocks more gameplay content. (Wikipedia).
We maximised every slider – FXAA was enabled, although we left all other Anti Aliasing settings disabled – based on reader feedback from previous reviews. ‘Ignore Suggested Limits’ was turned ‘ON’.
We found some intensive sections of the Grand Theft Auto 5 world and tested each card multiple times to confirm accuracy. The game demanded around 3.5GB of GPU memory at 1440p and just over 4GB at 4K.
Performance is excellent, averaging 64 frames per second at 1440p. That said, the reference GTX980ti clearly outperforms the AMD solution, by around an average of 8 frames per second. That said, in reality it is not that noticeable as the minimum frame rate via the Fury X 4GB is very consistent.
As was the case with The Witcher 3 at 1440P, the same point for the R9 Fury X minimum frame rate and a typical 1440P FreeSync monitor's lower boundary can be made. Using the above settings in GTA V, the R9 Fury X is the only single-GPU AMD card that can cleanly run in the general operating range for a typical 1440P FreeSync monitor.
Grand Theft Auto V is an action-adventure game played from either a first-person or third-person view. Players complete missions—linear scenarios with set objectives—to progress through the story.
Outside of missions, players can freely roam the open world. Composed of the San Andreas open countryside area and the fictional city of Los Santos, the world of Grand Theft Auto V is much larger in area than earlier entries in the series.
The world may be fully explored from the beginning of the game without restrictions, although story progress unlocks more gameplay content. (Wikipedia).
We maximised every slider – FXAA was enabled, although we left all other Anti Aliasing settings disabled – based on reader feedback from previous reviews. ‘Ignore Suggested Limits’ was turned ‘ON’.
We found some intensive sections of the Grand Theft Auto 5 world and tested each card multiple times to confirm accuracy. The game demanded around 3.5GB of GPU memory at 1440p and just over 4GB at 4K.
At Ultra HD 4K, the Fury X struggles to compete against the reference GTX980Ti even when manually overclocked to 1,130mhz. The Gigabyte GTX980 Ti G1 Gaming dominates the single GPU results, holding an average of 46 frames per second. The fastest card in Grand Theft Auto 5 at 4K however is still the AMD R9 295X2, by a clear margin.
On May 22, 2014, a Redux version of Metro Last Light was announced. It was released on August 26, 2014 in North America and August 29, 2014 in Europe for the PC, PlayStation 4 and Xbox One. Redux adds all the DLC and graphical improvements. A compilation package, titled Metro Redux, was released at the same time which includes Last Light and 2033. (Wikipedia). We test with following settings: Quality-Very High, SSAA-off, Texture Filtering-16x, Motion Blur-Normal, Tessellation-Normal, Advanced Physx-off.
We benchmark at some of the most intensive sections of the game, causing frame rates to dip. Many sections of the game are much smoother than this – so again, this is a strictly ‘worst case scenario'.
The Fury X holds onto the reference clocked GTX980 ti closely, although it is clearly outperformed at this resolution again, by around 5 frames per second. When we manually overclock the Fury X to 1,130mhz, the differential drops to around 2 frames per second with the all important minimum frame rate almost identical (1fps in favour of GTX980 Ti).
On May 22, 2014, a Redux version of Metro Last Light was announced. It was released on August 26, 2014 in North America and August 29, 2014 in Europe for the PC, PlayStation 4 and Xbox One. Redux adds all the DLC and graphical improvements. A compilation package, titled Metro Redux, was released at the same time which includes Last Light and 2033. (Wikipedia). We test with following settings: Quality- High, SSAA-off, Texture Filtering-16x, Motion Blur-Normal, Tessellation-Normal, Advanced Physx-off.
At Ultra HD 4K resolutions, the Fury X scores well, averaging 51 frames per second. When manually overclocked, the frame rates increase to 54 frames per second putting it ahead of the reference clocked Titan X. The Gigabyte GTX980 Ti G1 Gaming manages to dominate the average frame rate results, although the minimum frame rates are in the same zone.
We have pointed out the R9 Fury X card's minimum frame rate numbers and how they closely tie in with minimum FreeSync operating ranges of common 1440P monitors (using our image quality settings). Metro Last Light Redux extends that point to 4K – the R9 Fury X manages to post a minimum frame rate level of just over 40 FPS when overclocked, allowing it to run a typical 4K FreeSync monitor within the adaptive-syncing technology's operating range. No other single-GPU AMD card can manage something similar at these settings and resolution.
The tests were performed in a controlled air conditioned room with temperatures maintained at a constant 23c – a comfortable environment for the majority of people reading this.Idle temperatures were measured after sitting at the desktop for 30 minutes. Load measurements were acquired by playing Crysis Warhead for 30 minutes and measuring the peak temperature. We also have included Furmark results, recording maximum temperatures throughout a 30 minute stress test. All fan settings were left on automatic.
The cooler copes very well, with load temperatures hitting around 57c under gaming load. The cooler copes with the synthetic Furmark stress test, hiting 61c after sustained load durations.
We install the graphics card into our system and measure temperatures on the back of the PCB with our Fluke Visual IR Thermometer/Infrared Thermal Camera. This is a real world running environment.
Details shown below.
The card does get very warm to the touch after a little gaming and after 90 minutes gaming we measured 56.5C on the rear of the outer shell.
We noticed that the Fury X radiator was expelling very hot air out the rear of our case so we pointed the thermal imaging camera at the middle of the radiator grill and recorded a temperature of almost 55C. This was after 30 minutes of gaming.
From a thermodynamics and heat transfer perspective, the radiator's fins running hot allow ‘more efficient' (in essence) heat transfer than if they were running cooler (the principle of convective heat transfer). Users may be quick to point faults in a hot radiator, but this actually dissipates thermal energy to the surrounding air faster than a colder radiator would (all else being equal). A by-product of this is that AMD could, in theory, get away with running the fan at a lower speed while still achieving capable heat transfer rates. And as far as almost 55°C goes, the radiator's metal can handle that temperature without longevity concerns.
We have built a system inside a Lian Li chassis with no case fans and have used a fanless cooler on our CPU. The motherboard is also passively cooled. This gives us a build with almost completely passive cooling and it means we can measure noise of just the graphics card inside the system when we run looped 3dMark tests.
We measure from a distance of around 1 meter from the closed chassis and 4 foot from the ground to mirror a real world situation. Ambient noise in the room measures close to the limits of our sound meter at 28dBa. Why do this? Well this means we can eliminate secondary noise pollution in the test room and concentrate on only the video card. It also brings us slightly closer to industry standards, such as DIN 45635.
KitGuru noise guide
10dBA – Normal Breathing/Rustling Leaves
20-25dBA – Whisper
30dBA – High Quality Computer fan
40dBA – A Bubbling Brook, or a Refrigerator
50dBA – Normal Conversation
60dBA – Laughter
70dBA – Vacuum Cleaner or Hairdryer
80dBA – City Traffic or a Garbage Disposal
90dBA – Motorcycle or Lawnmower
100dBA – MP3 player at maximum output
110dBA – Orchestra
120dBA – Front row rock concert/Jet Engine
130dBA – Threshold of Pain
140dBA – Military Jet takeoff/Gunshot (close range)
160dBA – Instant Perforation of eardrum
The Fury X is reasonably quiet when idle, but there is pump noise evident at all times.
Additionally our sample did exhibit coil whine when gaming – and it seems linked to the frame rate. As an additional test I confirmed this was still apparent with both Superflower and Seasonic power supplies in the system.
Unfortunately this actually negatively impacted the dBa readings above for the Fury X – as between idle and load, the fan on the radiator was very quiet. Most of the noise generated was actually from the pump, and associated coil whine. I got used to the pump noise after a few hours, but the coil whine was quite irritating.
This does not appear to be isolated to our particular sample as readers have sent us video links such as this, this, this and this regarding Fury X noise issues.
It could be argued, however, that one of the benefits of using a radiator and fan that are physically separate from the graphics card (and connected via tubing) is noise control. What is meant by ‘control‘ is that the user has a degree of freedom as to where the radiator and fan are positioned. This may help a system builder control their noise levels if, for example, they have the graphics card mounted in the case that is close to them and then mount the radiator and fan in a separate position or chamber of the case (think of a chassis such as the CaseLabs Magnum M10, for example). The noise-emitting fan can be moved further away from a user's ears.
We measure power consumption from the whole system when idle and when gaming, excluding the monitor.
Power consumption has been improved when considering the last generation of AMD graphics cards -although the Fury X does still demand more power than the GTX980 Ti and Titan X.
The use of water-cooling and HBM both help AMD reduce the R9 Fury X power consumption to a level where it is similar to the R9 290X, but a considerably faster card.
The R9 Fury X 4GB is a positive move forward for AMD as it indicates the potential they have to produce small, powerful graphics cards – able to fit inside very small cases and system builds. HBM certainly seems to be tightly engrained into the future of graphics cards, even though the first iteration ships with a rather major caveat – the 4GB stacked limitation.
Right now in the real world, a 4GB framebuffer isn't really an issue if you want playable frame rates at 4K resolutions. I did notice many of our readers on Facebook were holding back from a GTX980 purchase claiming that 4GB isn't futureproofed. When the GTX980 Ti launched, that extra 2GB of memory seems to have appeased a wide audience. AMD, and the Fury X are likely to suffer from the same ‘4GB is just not enough' mindset.
Within the last week I have reviewed the AMD 390 (review HERE) and 390X (review HERE) , both of which ship with 8GB of GDDR5 memory. While undoubtedly reassuring to see on paper, it is clear to us that the 390 series Hawaii core runs out of horsepower at 4K before the 8GB of memory can be fully utilised.
While a head to head battle is going to very game specific, we can see that the R9 Fury X struggles to compete directly against Nvidia's GTX980 Ti. At 1440p the performance differential is heavily in favour of the GTX980 Ti, but at Ultra HD 4k the gap tightens considerably.
Most times the GTX980 Ti claims the top position, although the Fury X delivers fantastic results in Witcher 3, Grand Theft Auto 5 and Metro Last Light Redux at 4K resolutions and when manually overclocked it did manage to outperform the reference GTX980Ti in Metro LL.
Unfortunately for AMD, the reference GTX980ti is only part of the problem. There are many modified partner cards from MSI, Gigabyte and ASUS that bring substantial clock improvements to the table, as well as superlative cooling systems. Our results highlighted that the monster Gigabyte GTX980 Ti G1 Gaming literally leaves the R9 Fury X for dead, even when the AMD card is manually tweaked to the absolute limit.
Overclocking results from the R9 Fury X are a little disappointing. I had hoped that incorporating a liquid cooling system would reap dividends when we got to pushing the card past the reference speeds. I could only manage 80mhz extra from the Fiji core and while a software bug would allow memory sliders to be adjusted, it immediately caused blue screens on our test system.
One of my biggest concerns after testing the Fury X is related to the noise. There is no doubt that the Cooler Master ‘all in one' solution is proficient, and the fan has been designed to emit very little noise, even when loaded. Most of the noise our meter picked up was actually emitted from the pump and other related coil whine. As we detailed (with links) on the acoustics page, this doesn't seem to be isolated either which does raise some concerns. While we have heard that some early samples given to the press were at fault, our own sample is from the retail production line.
When I factor everything in, it is exciting to see AMD becoming competitive again in some areas. The R9 Fury X is not the GTX980 Ti killer that many people had expected and I do have some rather worrying concerns about the cooling system and coil whine. We are also disappointed to see that AMD have NOT decided to incorporate HDMI 2.0 support into the Fury X – meaning many people gaming on large televisions are stuck with 30hz at 4K. This is far from ideal.
The custom designed Nvidia GTX980 Ti partner cards are still our recommended high end choice.
I asked one of KitGuru's Senior Editors (Motherboards and Processors) Luke Hill to get a second opinion on The R9 Fury X:
While AMD's newest single-GPU flagship is not the GTX 980 Ti-killer many were hoping for, it does manage to restore the red team's competitiveness in the high-end GPU market. I feel that some of the technological innovations, especially the first widespread application of HBM, and continued use (leading on from the R9 295 X2) of an AIO water-cooling unit are positive points which lay foundations for the future of graphics cards.
As a fan of the benefits that monitors with synchronising technologies (G-Sync and FreeSync) add to a gaming experience, I feel that the R9 Fury X deserves some credit in this department. Many current 1440P FreeSync monitors seem to be hitting a sweet-spot with gamers, and some are considerably cheaper than similar G-Sync offerings. With many of the monitors featuring a lower boundary of 35 or 40 FPS for FreeSync to run, the R9 Fury X seems to be AMD's first single-GPU card that can consistently push minimum frame rates above the 35/40FPS FreeSync limits in many of today's games at their highest image settings. That, by many, will be deemed a positive from AMD's work.
However, from a raw performance standpoint, the R9 Fury X performance levels force it into a position where the card needs to be undercutting Nvidia's GTX 980 Ti in terms of pricing. A quick look at OverclockersUK shows a ~5% price benefit for the cheapest R9 Fury X against the cheapest GTX 980 Ti, which happens to be a reference model. That slight saving and the benefit of a water-cooling system may be enough to persuade some buyers, although it's quite an ask given the GTX 980 Ti's overall performance lead. When we start entering the £550+ region, AMD's flagship and its limited overclocking potential will struggle to compete against heavily-overclocked, custom-cooled GTX 980 Ti boards.
Both myself and my colleague Allan agree that if AMD can drop the price more towards £449.99 inc vat and get a fix for the coil whine and pump noise, we do feel the R9 Fury X has a logical place in the market.
You can buy AMD R9 Fury X partner cards from Overclockers UK HERE. Prices currently range between £509.99 and £649.99 inc vat.
Our American readers can get the Visiontek R9 Fury X 4GB sample from Amazon in the US for $682.17 with free shipping HERE.
Discuss on our Facebook page, over HERE.
Pros:
- runs cool under load.
- super quiet fan.
- physically small, ideal for specific system builds.
- HBM is likely the future of GPU design.
- AMD's greatest single GPU card yet.
- strong 4K performance.
Cons:
- pump noise.
- radiator takes up more space in case.
- coil whine is intrusive.
- minimal overclocking potential at this stage.
- the GTX980Ti is still the faster board.
- struggles to compete at 1440p.
- No HDMI 2.0 support.
Kitguru says: The R9 Fury X is a daring first step for AMD into the world of HBM. While on paper the card looks to be ‘all dominating', in the real world it struggles to compete against the Nvidia GTX980Ti. We also have some concerns over pump noise and associated coil whine.
