What's new
  • Please do not post any links until you have 3 posts as they will automatically be rejected to prevent SPAM. Many words are also blocked due to being used in SPAM Messages. Thanks!

AMD Radeon R9 290X 4GB Review

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,410
Location
Montreal
Dirt: Showdown / Far Cry 3

Dirt: Showdown (DX11)


<iframe width="560" height="315" src="http://www.youtube.com/embed/IFeuOhk14h0?rel=0" frameborder="0" allowfullscreen></iframe>​

Among racing games, Dirt: Showdown is somewhat unique since it deals with demolition-derby type racing where the player is actually rewarded for wrecking other cars. It is also one of the many titles which falls under the Gaming Evolved umbrella so the development team has worked hard with AMD to implement DX11 features. In this case, we set up a custom 1-lap circuit using the in-game benchmark tool within the Nevada level.


2560 x1440





Far Cry 3 (DX11)


<iframe width="560" height="315" src="http://www.youtube.com/embed/mGvwWHzn6qY?rel=0" frameborder="0" allowfullscreen></iframe>​

One of the best looking games in recent memory, Far Cry 3 has the capability to bring even the fastest systems to their knees. Its use of nearly the entire repertoire of DX11’s tricks may come at a high cost but with the proper GPU, the visuals will be absolutely stunning.

To benchmark Far Cry 3, we used a typical run-through which includes several in-game environments such as a jungle, in-vehicle and in-town areas.



2560 x1440


 
Last edited:

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,410
Location
Montreal
Hitman Absolution / Max Payne 3

Hitman Absolution (DX11)


<iframe width="560" height="315" src="http://www.youtube.com/embed/8UXx0gbkUl0?rel=0" frameborder="0" allowfullscreen></iframe>​

Hitman is arguably one of the most popular FPS (first person “sneaking”) franchises around and this time around Agent 47 goes rogue so mayhem soon follows. Our benchmark sequence is taken from the beginning of the Terminus level which is one of the most graphically-intensive areas of the entire game. It features an environment virtually bathed in rain and puddles making for numerous reflections and complicated lighting effects.


2560 x1440





Max Payne 3 (DX11)


<iframe width="560" height="315" src="http://www.youtube.com/embed/ZdiYTGHhG-k?rel=0" frameborder="0" allowfullscreen></iframe>​

When Rockstar released Max Payne 3, it quickly became known as a resource hog and that isn’t surprising considering its top-shelf graphics quality. This benchmark sequence is taken from Chapter 2, Scene 14 and includes a run-through of a rooftop level featuring expansive views. Due to its random nature, combat is kept to a minimum so as to not overly impact the final result.


2560 x1440


 
Last edited:

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,410
Location
Montreal
Metro: Last Light / Tomb Raider

Metro: Last Light (DX11)


<iframe width="640" height="360" src="http://www.youtube.com/embed/40Rip9szroU" frameborder="0" allowfullscreen></iframe>​

The latest iteration of the Metro franchise once again sets high water marks for graphics fidelity and making use of advanced DX11 features. In this benchmark, we use the Torchling level which represents a scene you’ll be intimately familiar with after playing this game: a murky sewer underground.


2560 x1440




Tomb Raider (DX11)


<iframe width="560" height="315" src="http://www.youtube.com/embed/okFRgtsbPWE" frameborder="0" allowfullscreen></iframe>​

Tomb Raider is one of the most iconic brands in PC gaming and this iteration brings Lara Croft back in DX11 glory. This happens to not only be one of the most popular games around but it is also one of the best looking by using the entire bag of DX11 tricks to properly deliver an atmospheric gaming experience.

In this run-through we use a section of the Shanty Town level. While it may not represent the caves, tunnels and tombs of many other levels, it is one of the most demanding sequences in Tomb Raider.


2560 x1440


 
Last edited:

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,410
Location
Montreal
Onscreen Frame Times w/FCAT

Onscreen Frame Times w/FCAT


When capturing output frames in real-time, there are a number of eccentricities which wouldn’t normally be picked up by FRAPS but are nonetheless important to take into account. For example, some graphics solutions can either partially display a frame or drop it altogether. While both situations may sound horrible, these so-called “runts” and dropped frames will be completely invisible to someone sitting in front of a monitor. However, since these are counted by its software as full frames, FRAPS tends to factor them into the equation nonetheless, potentially giving results that don’t reflect what’s actually being displayed.

With certain frame types being non-threatening to the overall gaming experience, we’re presented with a simple question: should the fine-grain details of these invisible runts and dropped frames be displayed outright or should we show a more realistic representation of what you’ll see on the screen? Since Hardware Canucks is striving to evaluate cards based upon and end-user experience rather than from a purely scientific standpoint, we decided on the latter of these two methods.

With this in mind, we’ve used the FCAT tools to add the timing of partially rendered frames to the latency of successive frames. Dropped frames meanwhile are ignored as their value is zero. This provides a more realistic snapshot of visible fluidity.





 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,410
Location
Montreal
Onscreen Frame Times w/FCAT (pg.2)

Onscreen Frame Times w/FCAT (pg.2)


When capturing output frames in real-time, there are a number of eccentricities which wouldn’t normally be picked up by FRAPS but are nonetheless important to take into account. For example, some graphics solutions can either partially display a frame or drop it altogether. While both situations may sound horrible, these so-called “runts” and dropped frames will be completely invisible to someone sitting in front of a monitor. However, since these are counted by its software as full frames, FRAPS tends to factor them into the equation nonetheless, potentially giving results that don’t reflect what’s actually being displayed.

With certain frame types being non-threatening to the overall gaming experience, we’re presented with a simple question: should the fine-grain details of these invisible runts and dropped frames be displayed outright or should we show a more realistic representation of what you’ll see on the screen? Since Hardware Canucks is striving to evaluate cards based upon and end-user experience rather than from a purely scientific standpoint, we decided on the latter of these two methods. With this in mind, we’ve used the FCAT tools to add the timing of runted to the latency of successive frames. Dropped frames meanwhile are ignored as their value is zero. This provides a more realistic snapshot of visible fluidity.





 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,410
Location
Montreal
Temperatures & Acoustics / Power Consumption

Temperature Analysis


For all temperature testing, the cards were placed on an open test bench with a single 120mm 1200RPM fan placed ~8” away from the heatsink. The ambient temperature was kept at a constant 22°C (+/- 0.5°C). If the ambient temperatures rose above 23°C at any time throughout the test, all benchmarking was stopped..

For Idle tests, we let the system idle at the Windows 7 desktop for 15 minutes and recorded the peak temperature.



Looking at the R9 290X’s temperatures, it feels like we’re back in the days of NVIDIA’s GTX 480 which was a historically hot-running piece of silicon. There is one difference though: according to AMD these sky-high temperatures are wholly intentional since they’re trying to wring every last ounce of performance from the core. This is done through the application of PowerTune’s new algorithms and it can be completely modified by changing the card’s temperature target. Naturally, this will lead to lower performance since the core won’t have as much thermal overhead.

Even the Uber Mode, which increases fan speeds to 55% exhibits the same behavior but behind the scenes, there are some differences. First of all, switching to Uber Mode doesn’t increase the power limits, temperature target or anything else but it does allow for some extra clock speed headroom since heat is being removed from the core at a higher pace, resulting in it requiring slightly less power. This means the temperature will remain at 94°C but average frequencies will be higher in any power-bound application.


Acoustical Testing


What you see below are the baseline idle dB(A) results attained for a relatively quiet open-case system (specs are in the Methodology section) sans GPU along with the attained results for each individual card in idle and load scenarios. The meter we use has been calibrated and is placed at seated ear-level exactly 12” away from the GPU’s fan. For the load scenarios, a loop of Unigine Valley is used in order to generate a constant load on the GPU(s) over the course of 15 minutes.


There are two different ways to look at these results. Silent Mode which is louder than comparable NVIDIA cards and Uber mode which is louder than anything we’ve tested in the last 18 months or so. Either way, the R9 290X isn’t a quiet card but in Silent Mode it’s doubtful you’ll hear it over 120mm case fans.


System Power Consumption


For this test we hooked up our power supply to a UPM power meter that will log the power consumption of the whole system twice every second. In order to stress the GPU as much as possible we used 15 minutes of Unigine Valley running on a loop while letting the card sit at a stable Windows desktop for 15 minutes to determine the peak idle power consumption.

Please note that after extensive testing, we have found that simply plugging in a power meter to a wall outlet or UPS will NOT give you accurate power consumption numbers due to slight changes in the input voltage. Thus we use a Tripp-Lite 1800W line conditioner between the 120V outlet and the power meter.


With a running temperature of 94°C, a 28nm manufacturing process that isn’t known for its efficiency and a huge number of transistors, the R9 290X consumes an epic amount of power for a single core card. In fact, it chugs down some 54W more than NVIDIA’s similarly-performing TITAN, 67W more than a GTX 780 and nearly as much as a GTX 690. Yes, THAT GTX 690.

Those are both worrying metrics for anyone who values lower power consumption but we do have to remember that this is an enthusiast-centric product that’s meant to be installed in high end systems with suitable power supplies. Is that an excuse? No, but it does give some context since we don’t expect the R9 290X to be installed into mid-range systems.
 
Last edited:

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,410
Location
Montreal
Overclocking the R9 290X Explained & Results

Overclocking the R9 290X Explained & Results


In the realm of enthusiast graphics cards, a lot of focus is put upon a product’s ability to go above and beyond the call of duty. This means plenty of overclocking headroom for anyone that wants a quick, easy performance uplift without having to look at higher end solutions. For the R9 290X, AMD has taken a completely new approach to this time-honored tradition.


As we’ve explained on the previous pages, AMD’s PowerTune plays a huge part in the R9 290X’s performance since it balances out numerous different aspects in an effort to maximize performance. Judging from the temperature results, it’s also evident that the card is operating very close to its limits which, once again, is due to PowerTune’s effective management of its domain.

Since engine clocks, temperature and the board’s power limit are all tied closely at the hip, the folks on the Radeon team have decided that frequency overclocking alone is too one-dimensional. To expand upon that, they’ve moved away from the standardized clock speed sliders we’ve all been using and have instituted a form of percentage-based overclocking which ties the power limit to core speed. Essentially, both have to be modified at the same time with the final overclock ultimately being determined by PowerTune’s algorithms. This means a set frequency will only be achieved if PowerTune sees enough thermal and power overhead. If neither of those factors is deemed to be sufficient for a set frequency, clock speeds will be automatically dialed back accordingly.

Earlier in the review we talk about the capability to personalize the gaming experience offered by the R9 290X and the multicolored chart above is where that begins. With it, a gamer can set higher speeds and power limits for a performance boost or, if a game is performing well above their expectations, dial things back so the board uses less power and runs quieter. Alternately, the fan speed and Temperature Target sliders can be used in tandem to essentially cap performance at a given thermal limit.

The main issue we see is that AMD’s tools have no way of telling users exactly what effect their changes will have upon actual core frequencies regardless of their relationship to other aspects within PowerTune. The problem facing AMD is quite simple: with so many aspects of performance tied at the hip to power limits, temperatures and whatnot, it’s nearly impossible to predict the effect varied applications will have upon an requested overclock. We can respect that since we’ve seen many supposedly “stable” Kepler overclocks where a peak number was reported but never continually achieved since Boost had to dial things back to ensure the ASIC operated within its preset boundaries. So AMD erred on the side of caution and is using what boils down to best-case percentages. Set them, keep a close eye on things and cross your fingers.


At first this whole process may seem a bit contrived but it actually works quite well in both Silent and Uber modes once you get the hang of it. As a rule of thumb, we moved towards the higher end of the spectrum (that’s what overclocks are for, right?) but the end results weren’t all that great. Even with fan speeds set to a screaming 70% the core speed hit a constant 1115MHz which isn’t all that bad and GDDR5 clocks hit 5684MHz before plateauing.

Unfortunately, when the RPMs were dialed down to a slightly more pedestrian 55%, overclocking was next to impossible with PowerTune constantly clamping down and returning frequencies to their default values after a few minutes. The R9 290X obviously rewards cooler temperatures with higher performance and that bodes well for board partners’ versions (which are due out before the end of November).

In reality, PowerTune’s new behavior may favor overclockers but may also allow them to post completely unrealistic results. Due to the lack of sustained load in programs like 3DMark, ultra high frequencies are ultimately quite achievable since the core will only be operating at its maximum frequency for short periods of time, followed by a cool-down period as the next benchmark loads. This allows the R9 290X to neatly sidestep the thermal limitations that start popping up as it remains under load. While our results are considered fully “gaming stable”, don’t expect every overclocking achievement to adhere to the same set of rules.

With all of that in mind, we don’t believe the R9 290X is a particularly good overclocker from a gamer’s perspective since its achievable, long-term clock speeds are typically lower than what you may have set.
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,410
Location
Montreal
Conclusion; A Titan Slayer Is Born

Conclusion; A Titan Slayer Is Born


AMD may have taken a while to get to this point but with the R9 290X, they finally have a viable competitor against NVIDIA’s GTX 780 and TITAN. More importantly, this new card is being introduced at a price that seriously undercuts the competition and ushers in a new day for the cost of high end GPUs. Is the R9 290X the stake to NVIDIA’s heart that so many AMD fans have been desperately hoping for? Only time will tell but expect it to cause a serious shakeup within the GeForce lineup in the near future.

AMD’s engineers have borrowed a page out of NVIDIA’s book by maximizing the Hawaii core’s area and transistor count in an effort to squeeze as much performance out of the 28nm manufacturing process as possible. They’ve gone about this in a logical way by optimizing the space allocated to the memory controllers but some major sacrifices were still made when creating the R9 290X. Double Precision performance operates at 1/8th speed while both heat production and power consumption are astronomical. It significantly outpaces the TITAN with a constant 95°C operating temperature and power draw that’s more than 50W higher.

AMD states that such high continual temperatures won’t affect ASIC longevity but numbers like that nonetheless point to a core being pushed to its limits. Our constrained overclocking results seem to back this hypothesis up as well.

The R9 290X’s temperature situation has been both mitigated and optimized through an updated version of PowerTune along with some additional on-die hardware elements. These work in tandem to balance out the core’s power needs and thermal output while also optimizing performance. In theory this approach pays off by insuring clock speeds remain at their highest possible levels. We did however see some odd but completely understandable results that caused us to completely rethink our benchmarking process.

As the R9 290X progresses through a gameplay scenario, its performance tends to drop by up to 7% as the PowerTune algorithms come to grips with stratospheric operating temperatures and the resulting increase in core power draw. This means the first two to five minutes of gameplay will see higher performance while frequencies tend to level out thereafter. NVIDIA’s similar GeForce Boost feature never encountered this kind of situation but a 10 minute “warm up” period was still necessary for all cards before every benchmark to replicate real-world gaming conditions.


PowerTune’s new functionality leads to a lot more personalization as well. This approach is personified by the R9 290X’s Uber and Silent modes which are selectable through a BIOS switch. Uber enhances performance by offering up acoustics as a sacrificial lamb while Silent balances out fan speeds and frequencies. Both can be further modified with clock speed tolerances via AMD’s slightly confusing but oh-so-useful Overdrive chart so there are plenty of options at your fingertips to make the R9 290X behave the way you want.

Even though the Radeon R9 290X is hugely powerful regardless of its preset, there are two different performance metrics we need to look at: those attained through Silent and Uber modes. With Silent Mode selected, it trades blows with NVIDIA’s TITAN and outpaces the GTX 780. There were some massive increases over the GTX 780 in games like Max Payne 3 and Hitman, both of which can make good use of the extra 1GB of memory on AMD’s card when played at extremely high settings.

The aptly named Uber Mode really moves performance to the next level. Through the application of increased fan speeds (55% versus Silent Mode’s 40%), the core is able to reach consistently higher frequencies. So, rather than fluctuating between 850MHz and 890MHz it steadies out around 980MHz. The results of this enhancement speak for themselves: on average, 6% better performance than Silent Mode places the R9 290X in truly elite company. At the very least, this gives AMD some serious bragging rights and it should also give NVIDIA some nightmares about what the upcoming R9 290 will bring to the table.


If the performance metrics of AMD’s R9 290X seem impressive, its price of $549 is quite simply a game changer. Not only does it make NVIDIA’s $649 GTX 780 look highly overpriced but the $999 TITAN has now been completely marginalized for anyone but entry level CUDA developers. The R9 280X’s launch at $299 should have put the writing on the wall but NVIDIA has been dreadfully slow to react to AMD’s new pricing structure. Sure, the bundling of free triple-A games with their GPUs does add some value. However, we can’t help but be highly critical of a strategy that ends up costing gamers so much more initially despite some great included titles. While reducing the cost of some lower end GeForce SKUs fired a shot across AMD’s bow, in one fell swoop AMD has just torpedoed NVIDA’s entire high end lineup.

The R9 290X can be as loud or quiet as you want but in its reference form, in order to achieve the performance numbers in this review, it is just massively louder than NVIDIA’s GTX 780 and TITAN. That’s a major problem for any gamers, particularly considering there won’t be any custom cooled versions available at launch from the likes of ASUS, MSI, Sapphire and XFX. Is the sacrifice worth it? Throw on some headphones and you’ll be none the wiser or wait for upgraded heatsinks / water blocks since the R9 290X really can fly when running at lower temperatures.

Some intangibles have been added into this mix too. TrueAudio and Mantle both sound like excellent concepts but they’re just too new for us to determine their true “value” as part of the R9 290X’s feature set. Then there’s NVIDIA’s GTX 780 Ti which, we would assume, will be (hopefully!) launched amid additional price drops for the now-overpriced GTX 770, GTX 780 and TITAN. Judging from NVIDIA’s direction for the GTX 700-series it may prove to be an excellent competitor while also being quieter and more efficient.

With the R9 290X, AMD has achieved something phenomenal. They have made extreme performance all that much more accessible to gamers by releasing a card that not only bucks current industry pricing trends but also looks towards the future with promising technologies Mantle and TrueAudio. Despite the fact that acoustics and power consumption may drag down the reference version in some respects, the R9 290X is a compelling enthusiast graphics card that has every right to brag about its supremacy.

 
Last edited:

Twitter

Top