The AMD R9 290 launched with a rather poorly designed stock cooling solution that let the card hit temperatures of 95°C under load! It seemed as if AMD was now targeting the ‘Room Heater’ market. A better Third party cooling solution was needed and we will find out over the course of this review if the Asus R9 290 DirectCU II OC provides one.
The Asus card comes not only equipped with a custom designed cooling solution, but also with a factory overclocked variant that was used for the purposes of this review. The overclock is slight with the Core Clock pushed from 947MHz to 1000 MHz and Memory Clock pushed from 1250MHz (5000 MHz effective) to 1260MHz (5040 MHz effective) which leaves room for additional overclocking.
[divider]
Specification of the R9 290 DirectCU II OC
The dimensions of the two slot card are roughly 28.7cm x 14.73cm x4.06cm and it features a custom Asus DirectCU II cooler with two 92mm fans, each with a distinctive look. The metal covering is adorned with three fierce looking red stripes. The card is rather wide and also thicker than other two slot cards. The only lights on the card are the green LEDs of the power connectors which are in a 8pin + 6pin configuration. The rear panel has two DVI-D connectors, one HDMI connector and one DisplayPort connector. The top of the card is also covered by a heat sheath.
The launch price for the R9 290 DirectCU II OC is ₹42,000 which should have dropped since the card’s release in January 2014.
Build Quality and Packaging
The build quality of the card is top notch and the card shows the attention to quality Asus has paid in its construction. The brushed metal heat sheath looks quite refined on the matte black PCB. The card has no Crossfire connectors as AMD has developed technology to allow on-board crossfire that works without the bridge. There is however a tiny switch almost hidden away on the card, that lets you switch it between quiet mode and maximum performance mode. We ran all our tests with the latter.
The card comes neatly packaged with multiple layers of anti-static plastic, Styrofoam and cardboard encasing it. It comes with the Driver CD, Quick installation Guide and a 2x6pin to 1x8pin PCI-E power adapter in their own discrete cardboard box. Plastic covers for the DVI connectors and the PCI-E pins add a neat touch.
[divider]
Testing Methodology
We tested the R9 290 DIRECTCU II OC not only for the average Frames per Second(FPS) but also for the 99th Percentile Frame time which tells us about the performance of the GPU within the second. Within the second testing is useful to understand micro-stutter which can render a game unplayable despite FPS being high. Fraps 3.5.99 allowed us to calculate both.
Since the card is a Top End GPU, we decided to use highest possible settings in our benchmark games. We tried to disable CPU dependent settings or minimise their impact where possible. VSync and frame buffering were disabled for testing. All tests were run at 1920×1080 on a single monitor configuration.
Test System:
CPU: AMD Phenom II X4 965 BE C3 @ 3.8GHz
Motherboard: ASRock 970 Extreme 3
RAM: 2x4GB G.Skill Sniper CL9 1600MHz, 2x2GB G.Skill Ripjaws CL9 1600MHz
PSU: Corsair TX650 650W (for MARS 760), Seasonic S12II Bronze 620W (for others)
HDD: 2xSeagate Barracuda 7200.12 1TB, 1xWesten Digital Red 3TB
OS: Windows 8.1 x64
NVIDIA Driver: 334.67 (Beta Driver)
AMD Driver: 14.3 Beta 1 (Beta Driver)
While this system may not look like a purpose built test rig, we decided to use a normal usage PC so as to better reflect real world scores of the card. The games were also tested with a few applications like Antivirus, Browser, VOIP tool and others running in the background to get a realistic usage scenario. All HDDs were thoroughly defragmented prior to usage. Due to lack of equipment we were unable to conduct acoustic and power testing.
3DMark 11 Firestrike
3Dmark is an artificial Benchmarking tool whose Firestrike Test is very thorough on DirectX 11cards powering High-end PCs. The full test run for Firestrike includes 2 GPU only tests, a CPU dependent Physics Test and a Combined Graphics and Physics Test. The Tool is also useful for stress testing a GPU when run on loop.
Given that we are looking for the Performance of the Card itself, one should look at the Graphics score and the FPS for Graphics tests 1 and 2. The Physics and Combined tests are CPU dependent which is the limiting factor of our test rig.
Batman Arkham Origins
Batman Arkham Origins is a game that’s been supported by NVIDIA and utilises their PhysX technology to handle physics. We used highest possible settings. PhysX was also set to Enhanced for Nvidia cards and turned off for AMD cards making a direct comparison impractical. We ran the inbuilt benchmark tool with a FRAPS timed run of 120 seconds.
Turning PhysX off seems to almost double the FPS of the game. However since we used this game to benchmark PhysX performance among NVIDIA cards, these results can be ignored. If PhysX were turned on, the performance becomes CPU dependent for the AMD card and drops much lower. The lower frametime is also consistent with the increase in FPS.
The 99th Percentile graph has a few odd spikes which might be interpreted as microstuttering in the game. However these are few and far between as to not render the game unplayable and are likely caused by area transitions in the benchmark scenes. The rest of the graph shows much smaller spikes and gameplay appeared to be quite fluid.
Battlefield 4
Battlefield 4 uses the Frostbite 3 engine to push the visual processing boundaries of current hardware. We ran the DX11 version with the highest possible settings and we also tested AMD c cards with Mantle as you will see later. The game offers no benchmark tool, so we used areas from the first single player campaign mission for the FRAPS run.
The R9 290 DirectCU II OC provides healthy competition to the GTX 770 but gets beaten by the latter’s lower 99th percentile frame time. The MARS being a dual GPU solution muscles ahead of the two.
The tight frame time graph is indicative of smooth game performance with no immersion breaking microstuttering. The run shows a consistent FPS for most of the test and the frame time graph shows no odd variances as well.
Bioshock Infinite
The Unreal Engine 3 powered Bioshock Infinite really pushed the boundaries of visual effects achievable with the ageing engine. It comes with a built in benchmark mode which runs for a lower time period than the standard 120s we used in other tests.
As expected, the R9 290 DirectCU II OC stays ahead of the GTX 770 but is soundly beaten by the MARS. However, it comes out as the highest in the 99th percentile frametime, though all the cards are very close.
The frame time graph is relatively tight save for the one spike which might have occurred during a transition between benchmark scene areas. The run was fluid with no apparent microstuttering though the fps drops did draw notice.
Company of Heroes 2
Relic’s Company of Heroes 2 is a tough nut to crack for quite a few GPUs, though it’s dependent on CPUs to a great degree as well. We used highest possible Settings with Physics turned off for the short 45s benchmark run the game offers.
The R9 290 DirectCU II OC is surprisingly faster than both NVIDIA cards and has a significantly lower 99th Percentile frame time. The MARS 760 is hampered by the game failing to recognise it as a SLI configuration and treating it as a single GTX 760 though. Curiously, the AMD card also has the lowest minimum FPS.
The Benchmark run was quite smooth as can be seen from the frame time graph, with the lone spike being attributable to a scene and area transition.
Crysis 3
CryEngine 3 was built to push the PC Hardware to its limits with Crysis 3 and deliver Maximum Eye Candy. We used the highest possible settings for the purpose of our test. The test run consisted of a section of the first mission for 120s as the game surprisingly offers no benchmark tool.
The R9 290 DIRECTCU II OC lets one comfortably play the game in all its graphical glory while the GTX 770 languishes in barely playable territory. The card does dip below 30fps at times but manages a lower 99th percentile frame time than both the GTX 770 and the MARS.
The frame time graph is not quite as smooth as the game during our test run. There are a lot of noticeable spikes and corresponding dips in FPS which can point to microstuttering being an issue. However, no major hiccups were noticeable and the run seemed smooth.
Far Cry 3
Far Cry 3 is a visually impressive game that lacks a benchmark mode. We used highest possible settings and the test run consisted of a 120s sample from the game’s single player campaign where we spent some time goofing off in the open world.
The R9 290 DIRECTCU II OC performs significantly better than the GTX 770 and the MARS in this game which hints at better optimisation for AMD cards. It also has a lower 99th percentile frame time and a higher minimum FPS.
The frametimes are closely distributed with no major spikes which is indicative of fluid gameplay with no noticeable stuttering. The game’s FPS does seem to fluctuate a lot though.
GRID 2
GRID 2 from Codemasters is one of the few games in our test suite that has no allegiance to either NVIDIA or AMD, which makes it a neutral candidate to better judge the GPU’s performance. We used highest possible Settings to run the benchmark tool offered by the game.
The AMD card loses out to both NVIDIA cards in the average FPS and the 99th percentile frametime which is curious in a vendor agnostic game.
The game’s frame time graph seems quite closely spaced though it fluctuates a lot. There are no big spikes and the gameplay was crisp and unblemished.
Sleeping Dogs
While the game is ageing, Sleeping Dogs is still a stunner when it comes to running benchmarks. We pushed the settings to highest possible and ran the game’s built in benchmark run.
The R9 290 DirectCU II OC soundly beats both NVIDIA cards in another AMD optimised game with higher average FPS and lower 99th Percentile frame time.
Transitions between scenes are the cause for frame time spikes and there was no microstuttering seen in the rendered scenes which are relatively smoother on the graph. Curiously there is a FPS drop in the closed spaces in interior areas while open world areas have a higher FPS.
Tomb Raider (2013)
Tomb Raider introduced us to a new Lara with fabled TressFX hair that behaves a lot more realistically than the pre-rendered mop we were used to. However the card does not seem to cope well with this technology and we had to drop it in our test run. We used highest possible settings with TressFX on. The Test run was shorter since the benchmark tool offered by the game runs for less than out 120s target time.
The R9 290 DIRECTCU II OC beats the GTX 770 by a large margin in this game, though it loses to the MARS. It has a 99th percentile frame time that’s roughly mid-way of those of the two NVIDIA cards. TressFX is a taxing technology but the AMD optimisation is apparent.
The frame time graph is closely spaced for the most part with few spikes at regular intervals indicative of microstuttering, which was noticeable at times.
Total War: Rome II
Total War: Rome II is another CPU heavy game that offers significant visual goodness. It offers a benchmark tool that focuses more on GPU power though and that is what we used for our test run of 120s. We set the game to Extreme and Unit Size to Small so as to reduce the impact of the CPU. Unlimited Video Memory was off so the game could scale down visual settings if it reached a bottleneck.
The R9 290 DirectCU II OC is very well suited to this game, beating the NVIDIA offerings on every parameter. The MARS is handicapped by the game recognising it as a single GTX 760 and not utilising the SLI configurations full power.
The frame times aren’t quite so closely spaced as we would like but there are only moderate spikes present. While the FPS does dip below 30, it is a rare occurrence and the game remains playable for most part.
[divider]
The Mantle Conundrum
AMD launched their new flagship GPUs with the announcement of a new API called Mantle that apparently allows low level hardware access and improves hardware performance in supporting games. While AMD claim the API is open and can be used by other vendors, the technology remains exclusive to their cards at the moment. Given that the same basic GPU units are also powering graphics on the new consoles, it seems sensible that AMD would want to pass on its benefits to consumers in the PC space with minimal hindrance to the game developers.
We decided to test and compare Mantle with DirectX 11 for Battlefield 4, Thief and the Star Swarm Benchmark. Since Mantle is not supported by FRAPS at the time of writing this review, we were forced to rely on the data generated by the applications themselves. Thief has an inbuilt benchmark tool that measures Average and Minimum FPS. However Frame time data is not exposed. On the other hand, Bttlefield 4 only has a basic tool to dump the frame time data which requires precise timing to attempt calculation of the average FPS. The Star Swarm Benchmark does allow generation of all the data parameters.
As can be seen, Mantle gives nearly a 33% boost in average FPS with an identical reduction in frame times. In a graphically demanding game like Battlefield 4, such a performance boost is quite significant.
While the numbers alone might make Mantle look good, the actual frame time plot paints a rather different picture. While the frame time is tightly concentrated for the most part, there are huge spikes at regular intervals which is indicative of noticeable microstuttering that can lead to an unplayable game. The DX 11 graph in comparison has a bigger distribution of spikes but much fewer large ones.
Thief seems to show about 30% increase in Average FPS with Mantle with a much higher minimum FPS. The game might dip into barely playable territory with DX11 but Mantle keeps it well above 30 FPS all the time.
Star Swarm boasts very high unit counts with individual AI for each unit using the GPU’s parallel processing abilities. The Benchmark is designed to showcase the abilities of Mantle and thus should only be used as a rough guideline. The scenes do look quite good when graphics are set to Extreme but it is the unit counts that are more impressive with about 5500 units onscreen at one time, something that has the potential to shake up the Real Time Strategy Genre.
We can see that the Mantle run is significantly faster with nearly double the average FPS and nearly half the 99th percentile frame time. These values are certainly impressive and show the biggest gain for mantle in our tests.
The Frame time graph comparison seems to suggest that as the unit counts increased towards the latter stages of the run, the DX11 version began to lag significantly to the Mantle one. The start of the benchmark seems to have large spikes for both runs that could be attributed to scene loading lag. The Mantle run shows similar regular spikes as in the battlefield 4 run with the rest of the graph being densely concentrated. However the spikes aren’t as pronounced as in Battlefield 4 and therefore result in no noticeable microstuttering. Strangely the benchmark also has similar spiking patterns in its DX11 run as well.
[divider]
Acoustics
We lack the equipment to conduct detailed acoustics testing on the card though we were able to get a rough idea of the card’s noise levels even at full load. Setting the fans at 100% makes your PC sound noticeably loud though in our load tests, the card barely exceeded 50% fan speed, which is much quieter. The card also offers a quiet mode with lower performance for the acoustically sensitive. We kept the case at 1m distance with the side panel open.
Temperatures
The Hawaii chips are quite hot like their namesake and the card would run at 46-48°C at idle even with the fans set to 100%. This was with an ambient temperature of 34°C. The card really shone in keeping its load temperature under check with a maximum of 81°C under full load for 15 minutes in our stress test. The fans barely went past 50% during this time which is indicative of the excellent cooler design by Asus. Given that the stock cooler lets the card run at temperatures of 95°C, this is a significant achievement and should be taken as a lesson by AMD when designing their next reference cooling solution.
Power
The R9 290 DIRECTCU II OC is rated at 250W though the card can exceed that when overclocked. The factory overclock is mild enough to stay within the power consumption limits of the stock GPU. The AMD card does have a power saving mode that dials down its frequencies and power requirement when idle.
Overclocking
The Asus R9 290 DirectCU II OC is a factory overclocked card, though given the relatively small bump, we wonder why they even bothered. Our answer comes from the overclocking performance of the card however. The Core clock can be pushed up to 15% with an increase in voltage to keep it stable while the memory clock refuses to be raised by more than 5%. Thus it seems the Asus Engineers themselves had little headroom to play with in releasing and Overclocked version. The extra overclocking can bring the card’s performance on par with a stock AMD R9 290X, which is quite good. While overclocking lets one extract maximum performance from a GPU, all of our test were conducted in the stock configuration.
Feature Set
Asus bundles its GPU tweak utility for easy overclocking of the card. AMD’s cards now come with support for their new proprietary technologies like Mantle and TrueAudio both of which seem to require special effort on the developers’ end to implement. While Mantle has support from a couple of released games with more games on the way and murmurs of greater developer support, Thief is the only game to support TrueAudio at the time of compiling this review. AMD has also tied up with Raptr to come out with the AMD Gaming Evolved app that suggests optimal game settings for one’s configuration, takes screenshots and videos, broadcasts to Twitch and tracks gametime. There’s even a rewards program in place to ensure adoption and even NVIDIA users can take advantage of this app.
[divider]
Conclusion
The Asus R9 290 DirectCU II OC is an excellent card when it comes to cooler design and build quality. It improves vastly over the drawbacks of the Stock AMD solution in the heat dissipation area whilst managing to remain tolerably quiet.
It is also a stellar performer when it comes to playing the latest games in all their visual fidelity. It seems especially adept with the strategy games in our tests. The card did have a strange quirk where in the entire diplay would corrupt upon Alt+Tabbing out of a game and then loading it up again, but this was easily fixed by Alt+Tabbing again.
The card’s exclusive technologies like Mantle and TrueAudio seem like a gimmick at the moment that only time will lift the curtains off of. The number of Mantle titles that can be tested is very small and new developments in DirectX seemingly reduce the gains promised by Mantle through greater efficiency. AMD’s trump card is that it currently powers all the major gaming consoles which might be interpreted as better developer support for AMD cards amongst multi-platform titles and widespread adoption of its technologies like Mantle by them.
While the card is factory overclocked and leaves a little headroom for further overclocking, we consider its asking price to be a little on the higher side. The craze with cryptocurrency mining that seems to favour AMD cards more might probably be attributed with keeping their prices high. However, we do recommend this card for anyone looking at getting the latest from the Red team while constrained by a budget. For its performance we award this card the iLL Gaming Silver Award.
We are extremely thankful to Asus for providing us with a test sample for reviewing.
What’s iLL
+Mantle and TrueAudio technologies hold future promise
+Excellent Cooler Design and Build Quality
+Performs quite well in the latest games with the highest settings
What’s Not
-Pricing is slightly on the higher side
-Quiet and Performance Mode switch is hard to locate
-Factory Overclock is slight and Overclocking headroom is small
