Original Link: http://www.anandtech.com/show/6025/radeon-hd-7970-ghz-edition-review-catching-up-to-gtx-680
AMD Radeon HD 7970 GHz Edition Review: Battling For The Performance Crownby Ryan Smith on June 22, 2012 12:01 AM EST
The GeForce GTX 680 launch marked both the closest and the farthest AMD has ever been from outright beating NVIDIA in modern times. On the one hand NVIDIA beat them by more than usual by achieving the holy trifecta as opposed to focusing just on performance. And yet on the other hand when it comes to raw performance AMD has never been this close. Where the GTX 580 beat the 6970 by 15% the GTX 680 led by just 10%, and even then it lost to the 7970 on some games. With such a close gap an obvious question arises: maybe, just maybe AMD could meet or beat NVIDIA with a higher clocked 7970 and rival them for the performance crown?
Today AMD is putting that idea to the test with the launch of the Radeon HD 7970 GHz Edition. Although AMD is not calling the 7970 GHz Edition a response to the GTX 680 – instead choosing to focus on it being Tahiti’s 6 month birthday – for all intents and purposes this is AMD’s response to the GTX 680. A higher clocked 7970 with AMD’s take on GPU turbo intended to make a run at the GTX 680 and that performance crown. So how does AMD fare? As we’ll see, after today it will no longer be clear who holds the performance crown.
Three months ago NVIDIA launched their GeForce GTX 680 to rave reviews and a boatload of editor recommendations, reclaiming their crown for the fastest single-GPU video card in the process. And for the first time in many years NVIDIA didn’t just beat AMD on raw performance, but they achieved the complete holy trifecta of video card competition – higher gaming performance, lower power consumption, and a lower price.
Consequently, for AMD this launch marked both the closest and the farthest they’ve ever been from outright beating NVIDIA in modern times. On the one hand NVIDIA beat them by more than usual by achieving the holy trifecta as opposed to focusing just on performance. And yet on the other hand when it comes to raw performance AMD has never been this close. Where the GTX 580 beat the 6970 by 15% the GTX 680 led by just 10%, and even then it lost to the 7970 on some games. With such a close gap an obvious question arises: maybe, just maybe AMD could meet or beat NVIDIA with a higher clocked 7970 and rival them for the performance crown?
Today AMD is putting that idea to the test with the launch of the Radeon HD 7970 GHz Edition. Although AMD is not calling the 7970 GHz Edition a response to the GTX 680 – instead choosing to focus on it being Tahiti’s 6 month birthday – for all intents and purposes this is AMD’s response to the GTX 680. A higher clocked 7970 with AMD’s take on GPU turbo intended to make a run at the GTX 680 and that performance crown. So how does AMD fare? As we’ll see, after today it will no longer be clear who holds the performance crown.
|AMD GPU Specification Comparison|
|AMD Radeon HD 7970 GHz Edition||AMD Radeon HD 7970||AMD Radeon HD 7950||AMD Radeon HD 6970|
|Memory Clock||6GHz GDDR5||5.5GHz GDDR5||5GHz GDDR5||5.5GHz GDDR5|
|Memory Bus Width||384-bit||384-bit||384-bit||256-bit|
|Manufacturing Process||TSMC 28nm||TSMC 28nm||TSMC 28nm||TSMC 40nm|
As far as performance and functionality goes, the Radeon HD 7970 GHz Edition (7970GE) is a rather straightforward upgrade to the existing Radeon HD 7970. In fact the hardware is absolutely identical right down to the GPU – there have been no changes to the PCB, the cooling, or the VRMs, and even the Tahiti GPU is the same revision that has been shipping in the 7970 since the beginning. Everything the 7970GE adds to the 7970 is accomplished through chip binning and new Catalyst and BIOS features specific to the 7970GE. So in many ways this is the 7970 we’ve already become familiar with, but with more pep in its step.
With identical hardware the real difference is in clockspeeds. The 7970 shipped at a rather conservative 925MHz core, which as we’ve seen in our 7970 overclocking adventures ends up being a good 175MHz less than what our worst 7970 can hit while overclocked without overvolting. At the time AMD left a lot on the table in order to maximize yields and to give their partners headroom to launch a range of factory overclocked cards, and now AMD has come to take that headroom back for themselves.
The 7970GE will ship at 1GHz, 75MHz faster than the 7970. Furthermore the 7970GE introduces AMD’s PowerTune Technology with Boost, which is AMD’s name for GPU turbo, and similar to the GPU turbo feature that is already on AMD’s APUs. The 7970GE can boost a further 50MHz up to 1050MHz, which means the 7970GE’s core clock increase is anywhere between 8% and 13.5% depending on how high it can go under a specific workload. We’ve seen that AMD’s performance scales very well with clockspeeds – which is to say it’s typically not memory bandwidth bottlenecked – so this bodes well for its performance. All the same AMD has also boosted their memory clocks from 5.5GHz to 6GHz, which will give the card 9% more memory bandwidth when it needs it. AMD hasn’t provided any specific guidance for performance, but overall you can expect around 10% better performance over the 7970 in GPU-bound situations, which is exactly what AMD needs to close the GTX 680 gap.
Beyond the higher clockspeeds and introduction of PowerTune Technology with Boost, that sums up the changes for the 7970GE. There are no board changes and it’s the same Tahiti GPU, meaning 2048 stream processors paired with 128 texture units and 32 ROPs, all on a 4.31B transistor GPU with a die size of 365mm2. With the increase in clockspeed from 7970 this pushes AMD’s theoretical double precision (FP64) compute performance over 1 TFLOPs to 1.08 TFLOPs, which AMD is in no way shy about mentioning since they’re the first GPU vendor to get there. On the memory side of things, AMD is using the same 3GB of GDDR5 we’ve previously seen, just clocked higher.
On that note, because AMD hasn’t made any hardware changes for the 7970GE the 7970GE’s TDP/PowerTune limit is equally unchanged. The 7970GE will have a PowerTune limit of 250W, identical to that of the 7970. With 6 months between the launch of the 7970 and the 7970GE, that’s 6 months of 28nm process improvements over at TSMC, which AMD will be using as the basis of their binning for the 7970GE. With that said there’s no such thing as a free lunch, and in practice the 7970GE’s power consumption has still increased relative to the 7970, as we’ll see in our benchmarks.
On a final point, at this point we would be remiss to not point out that once again AMD has once again added confusion to their product naming system in the name of simplicity. We have always pushed for clear naming schemes where parts with different specifications have different names and for good reason. AMD’s decision to name their new card the 7970 GHz Edition is unfortunate; while it’s true it has the same Tahiti GPU its performance and feature set (PowerTune Boost) are entirely different from the 7970. What’s the point of a 4 digit number if AMD is only ever going to use a fraction of them? In a rational universe this card would be the 7975 and that would be the end of that.
Our primary concern here is that a potential customer is going to read this review and then go out and buy a vanilla 7970 thinking they got the GHz Edition, which is the kind of misleading situation we want product names to avoid. At this point if AMD is going to continue producing multiple products under the name model number – and I can’t believe I’m saying this – they need to bring back proper suffixes. They were less sufferable than “GHz Edition”, which is just long enough to be ignored. At the end of the day clockspeed is not a proper product name.
Anyhow, with clocks and hardware settled, let’s talk about competitive positioning, pricing, and availability. As we alluded to in the introduction, the 7970GE is a clear swipe at the GeForce GTX 680. NVIDIA had a smaller than usual 10% lead with the GTX 680, and as a result AMD is making a run at it with a higher clocked Tahiti part. Realistically speaking, on average AMD can’t beat the GTX 680 with the 7970GE, but with good performance scaling they can tie.
Seeing as how it’s a GTX 680 competitor then, it should come as no surprise that AMD has put the MSRP on the 7970GE at $499, the exact same price as the GTX 680. It’s a slugfest for sure. At the same time it’s no secret that Tahiti cards are relatively expensive to manufacture – thanks to the larger-than-GK104 GPU and 3GB of GDDR5 – so AMD is keen on not just challenging NVIDIA for the crown but also bringing their margins back up to where they were prior to the GTX 680’s launch.
While the price of the 7970 and 7950 aren’t officially changing in the wake of the 7970GE’s launch, the launch of the GTX 600 series has already pushed pricing down to levels below even AMD’s April MSRPs. Reference clocked 7970s are down to around $430 after rebate, and the 7950 (having been pushed out of the picture by the GTX 670) is down to about $360 after rebate. Barring a move from NVIDIA, we expect AMD’s stack to settle here for the time being. As an aside, it looks like AMD will be continuing their Three For Free promotion for their existing 7900 series cards for some time to come, but they will not be extending it to the 7970GE. So while the 7970 will come with free games the 7970GE will not, which is going to further affect the value difference between the two cards.
Finally, while general card availability should be good – we’ve already seen that most 7970s can overclock to 7970GE speeds – AMD has pushed the launch out in front of when cards will actually ship. The 7970GE will not appear in stores until next week and widespread availability isn’t expected until July. But once cards do start flowing we don’t see any reason that AMD won’t be able to keep them in stock.
|Summer 2012 GPU Pricing Comparison|
|Radeon HD 7970 GHz Edition||$499||GeForce GTX 680|
|Radeon HD 7970||$429|
|$399||GeForce GTX 670|
|Radeon HD 7950||$359|
|Radeon HD 7870||$319|
|$279||GeForce GTX 570|
|Radeon HD 7850||$239|
Introducing PowerTune Technology With Boost
Since the 7970GE’s hardware is identical to the 7970, let’s jump straight into AMD’s software stack.
With the 7970GE AMD is introducing their own form of GPU Boost, which they are calling PowerTune Technology With Boost (PT Boost). PT Boost is a combination of BIOS and Catalyst driver changes that allow AMD to overdrive the GPU when conditions permit, and can be done without any hardware changes.
In practice PT Boost is very similar to NVIDIA’s GPU Boost. Both technologies are based around the concept of a base clock (or engine clock in AMD’s terminology) with a set voltage, and then one or more boost bins with an associated voltage that the GPU can move to as power/thermal conditions permit. In essence PT Boost allows the 7970GE to overvolt and overclock itself to a limited degree.
With that said there are some differences in implementation. First and foremost, AMD isn’t pushing the 7970GE nearly as far with PT Boost as NVIDIA has the GTX 680 with GPU Boost. The 7970GE’s boost clock is 1050MHz, a mere 50MHz better than the base clock, while the GTX 680 can boost upwards of 100MHz over its base clock. So long as both companies go down this path I expect we’ll see the boost clocks move higher and become more important with successive generations, just like what we’ve seen with Intel and their CPU turbo boost, but for the time being GPU turboing is going to be far shallower than what we’ve seen on the CPU.
At the same time however, while AMD isn’t pushing the 7970GE as hard as the GTX 680 they are being much more straightforward in what they guarantee – or as AMD likes to put it they’re being fully deterministic. Every 7970GE can hit 1050MHz and every 7970GE tops out at 1050MHz. This is as opposed to NVIDIA’s GPU Boost, where every card can hit at least the boost clock but there will be some variation in the top clock. No 7970GE will perform significantly better or worse than another on account of clockspeed, although chip-to-chip quality variation means that we should expect to see some trivial performance variation because of power consumption.
On that note it was interesting to see that because of their previous work with PowerTune AMD has far more granularity than NVIDIA when it comes to clockspeeds. GK104’s bins are 13MHz apart; we don’t have an accurate measure for AMD cards because there are so many bins between 1000MHz and 1050MHz that we can’t accurately count them. Nor for that matter does the 7970GE stick with any one bin for very long, as again thanks to PowerTune AMD can switch their clocks and voltages in a few milliseconds as opposed to the roughly 100ms it takes NVIDIA to do the same thing. To be frank in a desktop environment it’s not clear whether this is going to make a practical difference (we’re talking about moving less than 2% in the blink of an eye), but if this technology ever makes it to mobile a fast switching time would be essential to minimizing power consumption.
Such fast switching of course is a consequence of what AMD has done with their internal counters for PowerTune. As a reminder, for PowerTune AMD estimates their power consumption via internal counters that monitor GPU usage and calculate power consumption based on those factors, whereas NVIDIA simply monitors the power going into the GPU. The former is technically an estimation (albeit a precise one), while the latter is accurate but fairly slow, which is why AMD can switch clocks so much faster than NVIDIA can.
For the 7970GE AMD is further refining their PowerTune algorithms in order to account for PT Boost’s voltage changes and to further improve the accuracy of the algorithm. The big change here is that on top of their load based algorithm AMD is adding temperatures into the equation, via what they’re calling Digital Temperature Estimation (DTE). Like the existing PowerTune implementation, DTE is based on internal counters rather than an external sensor (i.e. a thermal diode), with AMD using their internal counters and knowledge about the cooling system to quickly estimate the GPU’s temperature similar to how they estimate power consumption, with a focus on estimating power in and heat out in order to calculate the temperature.
The end result of this is that by estimating the temperature AMD can now estimate the leakage of the chip (remember, leakage is a function of temperature), which allows them to more accurately estimate total power consumption. For previous products AMD has simply assumed the worst case scenario for leakage, which kept real power below AMD’s PowerTune limits but effectively overestimated power consumption. With DTE and the ability to calculate leakage AMD now has a better power estimate and can push their clocks just a bit higher as they can now tap into the headroom that their earlier overestimation left. This alone allows AMD to increase their PT Boost performance by 3-4%, relative to what it would be without DTE.
AMD actually has a longer explanation on how DTE works, and rather than describing it in detail we’ll simply reprint it.
DTE works as a deterministic model of temperature in a worst case environment, as to give us a better estimate of how much current the ASIC is leaking at any point in time. As a first order approximation, ASIC power is roughly a function of: dynamic_power(voltage, frequency) + static_power(temperature, voltage, leakage).
Traditional PowerTune implementations assume that the ASIC is running at a worst case junction temperature, and as such always overestimates the power contribution of leaked current. In reality, even at a worst case ambient temp (45C inlet to the fansink), the GPU will not be working at a worst case junction temperature. By using an estimation engine to better calculate the junction temp, we can reduce this overestimation in a deterministic manner, and hence allow the PowerTune architecture to deliver more of the budget towards dynamic power (i.e. frequency) which results in higher performance. As an end result, DTE is responsible for about 3-4% performance uplift vs the HD7970 GHz Edition with DTE disabled.
The DTE mechanism itself is an iterative differential model which works in the following manner. Starting from a set of initial conditions, the DTE algorithm calculates dTemp_ti/dt based on the inferred power consumption over a previous timeslice (is a function of voltage, workload/capacitance, freq, temp, leakage, etc), and the thermal capacitance of the fansink (function of fansink and T_delta). Simply put, we estimate the heat into the chip and the heat out of the chip at any given moment. Based on this differential relation, it’s easy to work back from your initial conditions and estimate Temp_ti, which is the temperature at any timeslice. A lot of work goes into tuning the parameters around thermal capacitance and power flux, but in the end, you have an algorithmic way to always provide benefit over the previous worst-case assumption, but also to guarantee that it will be representative of the entire population of parts in the market.
We could have easily done this through diode measurements, and used real temperature instead of digital temperature estimates…. But that would not be deterministic. Our current method with DTE guarantees that two parts off the shelf will perform the same way, and we enable end users to take advantage of their extra OC headroom on their parts through tools like Overdrive.
By tapping into this headroom however AMD has also increased their real power consumption at lower temperatures and leakages, which is why despite the identical PowerTune limits the 7970GE will consume more power than the 7970. We’ll get into the numbers in our benchmarking section, but it’s something to keep in mind for the time being.
Finally, on the end-user monitoring front we have some good news and some bad news. The bad news is that for the time being it’s not going to be possible to accurately monitor the real clockspeed of the 7970GE, either through AMD’s control panel or through 3rd party utilities such as GPU-Z. As it stands AMD is only exposing the base P-states but not the intermediate P-states, which goes back to the launch of the 7970 and is why we have never been able to tell if PowerTune throttling is active (unlike the 6900 series). So for the time being we have no idea what the actual (or even average) clockspeed of the 7970GE is. All we can see is whether it’s at its boost P-state – displayed as a 1050MHz core clock – or whether it’s been reduced to its P-state for its base clock, at which point the 7970GE will report 1000MHz.
The good news is that internally of course AMD can take much finer readings (something they made sure to show us at AFDS) and that they’ll finally be exposing these finer readings to 3rd party applications. Unfortunately they haven’t given us an expected date, but soon enough their API will be able to report the real clockspeed of the GPU, allowing users to see the full effects of both PT Boost and PT Throttle. It’s a long overdue change and we’re glad to see AMD is going to finally expose this data.
Dueling Drivers, A PR Do-Over, & The Test
Coinciding with the launch of the 7970GE, both AMD and NVIDIA have released new drivers this week. For AMD the 7970GE will be launching with the Catalyst 12.7 beta, while NVIDIA has released the 304.48 betas for the entire lineup. For both AMD and NVIDIA these are major performance drivers, with both groups seeing significant performance in a handful of games.
We’ll take a look at the full performance spectrum starting with our regular benchmarks on the next page, but first we wanted to quickly break out the performance impact of the driver changes for both the 7970 and the GTX 680 at 2560x1600.
The biggest gains for AMD and NVIDIA are seen in DiRT 3 and Shogun 2 respectively. For reasons we’re not fully sure of, AMD has managed to significantly improve their performance on DiRT 3 and turn the tables on what was previously a game favoring NVIDIA. At the same time however NVIDIA has finally fixed the Kepler performance bug that was introduced in the March update for Shogun 2, restoring performance to where it was before the update and pushing NVIDIA ahead of AMD once more.
Meanwhile AMD sees lesser gains in Batman, BF3, and Skyrim, while NVIDIA also gains on Batman but that’s about it. Altogether AMD’s latest driver update has pushed their average performance ahead slightly more than NVIDIA’s, which for the 7970 versus the GTX 680 means that the GTX 680 now leads by about 9% instead of 10%. It’s not nearly enough to change any recommendations, but the fact that AMD and NVIDIA have swapped leads in certain games is a reminder of how important drivers can be and how volatile performance leads are.
A PR Do-Over
Because the 7970GE is functionally identical to the 7970 save its higher clocks and PT Boost, we’re opting to use our limited time to focus solely on performance. With that said it was interesting to see that AMD is very much treating the 7970GE launch as a do-over for the 7900 series. In their press presentation they spent a fair bit of time focusing on various aspects of the 7900 series and GCN we’re already familiar with, such as compute performance, their Leo technology demo, and of course GCN.
None of this has any direct relevance to the performance of the 7970GE, but we thought it was worth mentioning since it gives a bit more insight into how AMD is approaching things. This is all we’re going to say on the matter, but we have reproduced a few of the press slides below for any of you that are interested.
|CPU:||Intel Core i7-3960X @ 4.3GHz|
|Motherboard:||EVGA X79 SLI|
|Chipset Drivers:||Intel 22.214.171.1242|
|Power Supply:||Antec True Power Quattro 1200|
|Hard Disk:||Samsung 470 (256GB)|
|Memory:||G.Skill Ripjaws DDR3-1867 4 x 4GB (8-10-9-26)|
|Case:||Thermaltake Spedo Advance|
AMD Radeon HD 5870
AMD Radeon HD 6970
AMD Radeon HD 7870
AMD Radeon HD 7950
AMD Radeon HD 7970
AMD Radeon HD 7970 GHz Edition
NVIDIA GeForce GTX 470
NVIDIA GeForce GTX 580
NVIDIA GeForce GTX 670
NVIDIA GeForce GTX 680
NVIDIA ForceWare 304.42 Beta
AMD Catalyst 12.7 Beta
|OS:||Windows 7 Ultimate 64-bit|
Kicking things off as always is Crysis: Warhead. It’s no longer the toughest game in our benchmark suite, but it’s still a technically complex game that has proven to be a very consistent benchmark. Thus even four years since the release of the original Crysis, “but can it run Crysis?” is still an important question, and the answer continues to be “no.” While we’re closer than ever, full Enthusiast settings at a 60fps is still beyond the grasp of a single-GPU card.
Since the launch of the GTX 680 it’s been clear that Crysis is a game that favors AMD’s products and this is no clearer than with the 7970GE. AMD was already handily beating the GTX 680 here, most likely due to the GTX 680’s more limited memory bandwidth – so the faster 7970GE widens that gap even further. The 7970GE is a full 25% faster than the GTX 680 here at 2560 and is extremely close to hitting 60fps at 1920, which given Crysis’s graphically demanding nature is quite incredible, and for all practical purposes puts the 7970GE in its own category. Obviously this is one of AMD’s best games, but it’s solid proof that the 7970GE can really trounce the GTX 680 in the right situation.
As for the 7970GE versus the 7970, this is a much more straightforward comparison. We aren’t seeing the full extent of the 7970GE’s clockspeed advantage over the 7970 here, but the 7970GE is still at the lower bounds of its theoretical performance advantage over its lower clocked sibling with a gain of 8% at 2560. The 7970GE is priced some 16% higher than the 7970 so the performance gains aren’t going to keep pace with the price increases, but this is nothing new for flagship cards.
When it comes to minimum framerates the 7970GE further expands its lead. It’s now 35% faster than the GTX 680 (and just short of 30fps) at 2560, which neatly wraps up the 7970GE’s domination in Crysis. Even its performance lead versus the 7970 improves, with the 7970GE increasing its lead to 13%. A year ago NVIDIA and AMD were roughly tied with Crysis, but now AMD has clearly made it their game. So can it run Crysis? Yes, and a lot better than the GTX 680 can.
Paired with Crysis as our second behemoth FPS is Metro: 2033. Metro gives up Crysis’ lush tropics and frozen wastelands for an underground experience, but even underground it can be quite brutal on GPUs, which is why it’s also our new benchmark of choice for looking at power/temperature/noise during a game. If its sequel due this year is anywhere near as GPU intensive then a single GPU may not be enough to run the game with every quality feature turned up.
Metro is another graphically intensive game and another game that has traditionally favored AMD’s 7000 series cards, albeit not to nearly the same degree as Crysis. Still, the 7970 could already surpass the GTX 680 here and the 7970GE adds to that lead. At 2560 the 7970GE takes a 12% lead, pushing past 40fps at 2560. Meanwhile AMD’s gains compared to the 7970 are less pronounced but it’s still enough for an 8% improvement. Like Crysis this is another game where AMD has a very clear performance advantage with the 7970GE.
For racing games our racer of choice continues to be DiRT, which is now in its 3rd iteration. Codemasters uses the same EGO engine between its DiRT, F1, and GRID series, so the performance of EGO has been relevant for a number of racing games over the years.
With AMD’s recent and significant performance gains under DiRT 3 due to their latest drivers, what was once an NVIDIA dominated game has flipped to an AMD-leading game before we even factor in the 7970GE. The 7970GE in turn only adds insult to injury, opening up an 18% lead at 2560. At this point framerates are so far above 60fps that it’s largely an academic difference, but 120Hz gamers will appreciate the difference. Otherwise we’ll be switching to DiRT: Showdown soon enough, and that should help to pull everyone’s framerate back down.
We should note that this is one of a few games where multi-monitor performance looks good across the board. Generally speaking having 2 high-end GPUs is still necessary for multi-monitor gaming (at least at comparably high quality), but in DiRT 3 even a single 7970GE can push past 71fps with ultra quality and 4x MSAA.
The minimum framerates largely reflect what we’ve already seen with the average framerates. DiRT 3 is already pretty consistent so the minimums aren’t much lower than the averages, which means at 2560 everything sails past 60fps with the 7970GE taking a clear lead. Even 5670 sees the 7970GE clearing 60fps.
Total War: Shogun 2
Total War: Shogun 2 is the latest installment of the long-running Total War series of turn based strategy games, and alongside Civilization V is notable for just how many units it can put on a screen at once. As it also turns out, it’s the single most punishing game in our benchmark suite (on higher end hardware at least).
Similar to what AMD did with improving their DiRT 3 performance, NVIDIA has similarly improved their Shogun 2 performance. NVIDIA was previously bitten by a major performance regression introduced in a March update for Shogun 2, which significantly reduced performance at 2560 with our Ultra settings. With NVIDIA having fixed that regression through a driver update, their performance has now been restored.
Consequently this is the first game where the 7970GE can’t power ahead of the GTX 680 at 2560. The 7970GE is just fast enough to break past 30fps, but compared to the GTX 680 the latter has a 12% performance lead here, which is a big deal for NVIDIA since performance is low across the board here.
Like we saw with the GTX 670 versus the 7970 and the GTX 570 versus the 6970, performance patterns are quickly devolving to the point where any performance leads are extremely inconsistent and strongly depend on the game in question. Or in other words, we’re going to see a lot of back-and-forth movement with the 7970GE winning with some games and the GTX 680 winning with others. For better or worse there won’t be a single clear winner across every game.
Batman: Arkham City
Batman: Arkham City is loosely based on Unreal Engine 3, while the DirectX 11 functionality was apparently developed in-house. With the addition of these features Batman is far more a GPU demanding game than its predecessor was, particularly with tessellation cranked up to high.
Previously both the 7970 and GTX 680 were tied here at 2560. With both AMD and NVIDIA improving their performance here however, things have finally begun to separate. The GTX 680 now has a small-but-measurable lead over the 7970, but it’s not enough to hold off the 7970GE. As a result the 7970GE enjoys a 5% performance lead at 2560. Strangely however it falls behind by a significant margin at 5760, which is not what we would have expected.
As a side note, it's worth mentioning that Batman is one of the titles with PhysX support that can actually make a difference in how the game looks. While performance with our non-PhysX settings may be close, there's something to be said for the PhysX effects here. That's an NVIDIA exclusive, and if it's something you want you'll have to go with an NVIDIA GPU.
Portal 2 continues the long and proud tradition of Valve’s in-house Source engine. While Source continues to be a DX9 engine, Valve has continued to upgrade it over the years to improve its quality, and combined with their choice of style you’d have a hard time telling it’s over 7 years old at this point. Consequently Portal 2’s performance does get rather high on high-end cards, but we have ways of fixing that…
The great thing about having such powerful cards is that we can push image quality to the max on better performing games, and in no place is that better evident than Portal 2. Here we can get away with 4x Super Sample Anti-Aliasing even at 2560, providing smooth and beautiful gameplay.
Unfortunately for AMD, they just don’t do very well on this test and the 7970GE doesn’t do much to remedy that. Where the GTX 680 can surpass 60fps the 7970GE can only manage 55fps, giving the GTX 680 a commanding 25% lead (or causing the 7970GE to trail by 20%). The 7970GE is more than playable here, but it would be nice to see it pass 60fps to maximize how fluid the game is. The situation of course improves without SSAA, but at these performance levels it doesn’t make a whole lot of sense not to use it at single-monitor resolutions.
Its popularity aside, Battlefield 3 may be the most interesting game in our benchmark suite for a single reason: it’s the first AAA DX10+ game. It’s been 5 years since the launch of the first DX10 GPUs, and 3 whole process node shrinks later we’re finally to the point where games are using DX10’s functionality as a baseline rather than an addition. Not surprisingly BF3 is one of the best looking games in our suite, but as with past Battlefield games that beauty comes with a high performance cost.
Battlefield 3 continues to be NVIDIA’s ace, but the 7970GE combined with some minor driver performance improvements AMD has eroded that lead some. At 2560 the GTX 680 enjoys a 11% lead over even the 7970GE and even the GTX 670 can overtake the 7970GE, but at the very least AMD is on the threshold of 60fps here. More optimistically they have improved their performance over the 7970 by 11%, which is an above average performance gain for the 7970GE. AMD will likely never close the gap on current hardware but it will be in their best interests to keep it narrow for future Frostbite Engine 2 games.
Our next game is Starcraft II, Blizzard’s 2010 RTS megahit. Much like Portal 2 it’s a DX9 game designed to run on a wide range of hardware so performance is quite peppy with most high-end cards, but it can still challenge a GPU when it needs to.
Starcraft II has traditionally favored NVIDIA and neither driver improvements nor higher clockspeeds can change that. The 7970GE closes the gap, but it still leaves the GTX 680 ahead by 3% at 2560. This is however a solid improvement over the 7970, with the 7970GE having increased its performance by 9% over its predecessor. With the next edition of Starcraft II still technically due this year it will be interesting to see if performance shifts at all or whether NVIDIA maintains their lead.
The Elder Scrolls V: Skyrim
Bethesda's epic sword & magic game The Elder Scrolls V: Skyrim is our RPG of choice for benchmarking. It's altogether a good CPU benchmark thanks to its complex scripting and AI, but it also can end up pushing a large number of fairly complex models and effects at once, especially with the addition of the high resolution texture pack.
For the longest time AMD has had various problems with Skyrim, ranging from negative Crossfire scaling to mediocre performance. We haven’t had a chance to test the former, but with Catalyst 12.7 they’ve definitely fixed the latter. At 2560 the performance of the 7970 has improved to the point where it’s solidly CPU limited just like its NVIDIA competition, though for some reason NVIDIA continues to handle this situation slightly better than AMD.
In any case because the 7970 is now CPU limited at 2560 the 7970GE doesn’t have any room to grow. This isn’t a hardware fault, this is just a reflection of how RPGs tend to be CPU limited before they’re GPU limited. We may have to look at running extra graphics effects (e.g. TrSSAA/AAA) to thin the herd in the future.
Our final game, Civilization 5, gives us an interesting look at things that other RTSes cannot match, with a much weaker focus on shading in the game world, and a much greater focus on creating the geometry needed to bring such a world to life. In doing so it uses a slew of DirectX 11 technologies, including tessellation for said geometry, driver command lists for reducing CPU overhead, and compute shaders for on-the-fly texture decompression.
Civilization V is another game that has previously gone to NVIDIA and has now been disrupted by the introduction of the 7970GE. At 2560 the GTX 680 and 7970GE are now in a dead heat, with both cards squeezing past 60fps. The 7970GE picks up 7% over the 7970, which is just enough to close the gap versus the GTX 680 here.
Shifting gears, as always our final set of performance benchmarks is a look at compute performance. As we saw with the launch of the GTX 680, Kepler (GK104) just doesn’t do very well here, thanks in part to NVIDIA stripping out a fair bit of compute hardware and memory bandwidth on GK104 in order to focus on gaming performance. OpenCL performance is particularly bad with NVIDIA almost completely ignoring it, but even DirectCompute performance often swings AMD’s way. This isn’t to say that GK104 doesn’t have its moments, but when it comes to compute it’s typically AMD’s time to shine.
Our first compute benchmark comes from Civilization V, which uses DirectCompute to decompress textures on the fly. Civ V includes a sub-benchmark that exclusively tests the speed of their texture decompression algorithm by repeatedly decompressing the textures required for one of the game’s leader scenes. Note that this is a DX11 DirectCompute benchmark.
The 7970 already had a significant lead in this benchmark thanks to AMD’s work on improving their DirectCompute performance, and the 7970GE extends it further. The most important factor of course is actual game performance – where the 7970GE and GTX 680 are tied – but this is clear software evidence of what we already know in hardware: that the 7970GE is far more potent at compute than the GTX 680 is.
Our next benchmark is SmallLuxGPU, the GPU ray tracing branch of the open source LuxRender renderer. We’re now using a development build from the version 2.0 branch, and we’ve moved on to a more complex scene that hopefully will provide a greater challenge to our GPUs.
Being an OpenCL title that NVIDIA isn’t taking any care to optimize for, the 7970GE simply blows the GTX 680 out of the water. It’s not even a contest here. Only one card family is even worth consideration for use here. However it’s interesting to note that the 7970GE’s performance improvement over the 7970 is a bit below average, with the 7970GE only picking up 6%. SLG does stress memory bandwidth and compute performance, but in all likelihood the 7970GE isn’t boosting as much here as it is under our gaming tests. Once AMD starts exposing real clockspeeds we’ll need to revisit this assumption.
For our next benchmark we’re looking at AESEncryptDecrypt, an OpenCL AES encryption routine that AES encrypts/decrypts an 8K x 8K pixel square image file. The results of this benchmark are the average time to encrypt the image over a number of iterations of the AES cypher.
While the 7970GE does improve upon the 7970’s already strong performance, we’re clearly reaching the point where the relatively long CPU/GPU transfer times over PCIe are taking their toll, explaining why the 7970GE could only shave off 5ms. This is actually an important point to make and is why APUs are so important to AMD’s GPU computing plans, but it also means that at a certain speed GPU performance ceases to matter.
Our fourth benchmark is once again looking at compute shader performance, this time through the Fluid simulation sample in the DirectX SDK. This program simulates the motion and interactions of a 16k particle fluid using a compute shader, with a choice of several different algorithms. In this case we’re using an (O)n^2 nearest neighbor method that is optimized by using shared memory to cache data.
In this final compute shader benchmark NVIDIA’s performance is actually quite respectable, leading to them besting the 7970. However the 7970GE provides just enough of a performance boost to push AMD ahead of NVIDIA here, giving AMD a solid majority of our standard compute benchmarks. Even when Kepler is faced with a favorable workload, it looks like GCN based 7970GE is capable of taking NVIDIA head-on.
Finally, we received a number of requests for some further compute benchmarking using some of the consumer programs AMD provided the press with for the Trinity launch. In particular WinZip and handbrake were requested, so we’ve gone ahead and run those benchmarks for this review.
Starting with WinZip, WinZip 16.5 introduced OpenCL acceleration of both compression and AES achieve encryption. Despite being accelerated via OpenCL WinZip only supports AMD devices, presumably because only AMD provided technical assistance. As a result we’re looking solely at pure CPU performance and GPU accelerated performance across AMD’s lineup.
One thing immediately sticks out: WinZip isn’t very sensitive to GPU performance. Merely having a GPU increases performance rather significantly, but it doesn’t matter if it’s a fast GCN card or a GCN card at all for that matter, as even the VLIW4 based 6970 returns the same times. In fact AMD’s drivers report almost no GPU load, so it’s questionable how much of this is actually being run on the GPU versus being run on the CPU through AMD’s OpenCL CPU driver.
As for Handbrake, AMD sent along a newer version that works with discrete GPUs. AMD notes that this is still very much a work in progress, which we saw first-hand when OpenCL acceleration failed to handle two of our three test clips. It failed to properly crop one video, and failed to properly detelecine another. Handbrake’s OpenCL acceleration will of course continue to improve as it approaches release, but for the time being it’s definitely a beta.
Much like WinZip, Handbrake doesn’t appear to be particularly GPU performance sensitive, which doesn’t come as much of a surprise. Large parts of the H.264 encoding process are ill suited for GPU acceleration, so X.264 is only offloading part of the process and the deciding factor is still CPU performance. The actual GPU load is very inconsistent, but generally tops out at around 40% usage.
The end result is nothing to sneeze at however. Whereas Handbrake averaged 25.6fps without GPU acceleration, with it performance increases by 24% to around 32fps. And unlike other GPU compute accelerated encoders the quality here is very consistent between the CPU and GPU paths (though GPU file size tends to be a bit larger), which means we’re retaining the same quality and customizability of Handbrake/x264 while gaining additional performance for free.
Despite the fact that this is an AMD backed initiative it’s interesting to see that Handbrake’s performance isn’t heavily reliant on the GPU being used. We would have assumed that Handbrake was only optimized for AMD’s GPUs at this point, and even if that’s the case NVIDIA’s GPUs are still fast enough to make up the difference. The fact that Handbrake performance with NVIDIA’s GPUs is a hair faster is not at all what we would have expected, but at the same time this is very beta quality software and is likely dependent on the clip being used, so we wouldn’t advise reading too much into this at this time.
Because the 7970GE uses the same architecture – and indeed same chip as the 7970 – there’s not much to report here with synthetic performance, but we’ll quickly run through the numbers.
Starting with 3DMark Pixel and Texel Fill, the 7970GE improves on the 7970 around as expected. Pixel fill is a bit low at only 4.5% though, which is unexpected since the 7970GE got a big boost in both its core clock and its memory clock.
Next up is tessellation. While at normal tessellation levels we’re CPU or otherwise throughput limited, at maximum tessellation levels the 7970GE further improves upon the 7970 by 13%, which means we’re likely seeing the full impact of PT Boost in action here. However at this performance level we’re also approaching the same throughput limits we saw with normal tessellation, which means that we either need a better test or GPUs are reaching the point where they have more than enough tessellation performance for future games.
Finally we have Unigine’s Heaven. The GTX 600 series has always done well here, and while the 7970GE improves on the 7970, the GTX 680 is still well ahead for what it’s worth.
Power, Temperature, & Noise
As always, we’re wrapping up our look at a video card’s stock performance with a look at power, temperature, and noise. Officially AMD is holding the 7970GE’s TDP and PowerTune limits at the same level they were at for the 7970 – 250W – however unofficially because of the higher voltages, higher clockspeeds, and Digital Temperature Estimation eating into the remaining power headroom, we’re expecting power usage to increase. The question then is “how much?”
|Radeon HD 7970 Series Voltages|
|Ref 7970GE Base Voltage||Ref 7970GE Boost Voltage||Ref 7970 Base Voltage|
Because of chip-to-chip variation, the load voltage of 7970 cards varies with the chip and how leaky it is. Short of a large sample size there’s no way to tell what the voltage of an average 7970 or 7970GE is, so we can only look at what we have.
Unlike the 7970, the 7970GE has two distinct voltages: a voltage for its base clock, and a higher voltage for its boost clock. For our 7970GE sample the base clock voltage is 1.162v, which is 0.013v lower than our reference 7970’s base clock voltage (load voltage). On the other hand our 7970GE’s boost clock voltage is 1.218, which is 0.056v higher than its base clock voltage and 0.043v higher than our reference 7970’s load voltage. In practice this means that even with chip-to-chip variation, we’d expect the 7970GE to consume a bit more power than the reference 7970 when it can boost, but equal to (or less) than the 7970 when it’s stuck at its base clock.
So how does this play out for power, temperature, and noise? Let’s find out.
Starting with idle power, because it’s the same GPU on the same board there are no surprises here. Idle power consumption is actually down by 2W at the wall, but in practice this is such a small difference that it is almost impossible to separate from other sources. Though we wouldn’t be surprised if improving TSMC yields combined with AMD’s binning meant that real power consumption has actually decreased a hair.
Similar to idle, long idle power consumption is also slightly down. NVIDIA doesn’t have anything to rival AMD’s ZeroCore Power technology, so the 7970CE is drawing a full 10W less at the wall, a difference that will become more pronounced when we compare SLI and CF in the future.
Moving on to our load power we finally see our first 7970GE power results, and while it’s not terrible it’s not great either. Power at the wall has definitely increased, with our testbed pulling 429W with the 7970GE versus 391 with the 7970. Now not all of this is due to the GPU – a certain percentage is the CPU getting to sleep less often because it needs to prepare more frames for the faster GPU – but in practice most of the difference is consumed (and exhausted) by the GPU. So the fact that the 7970GE is drawing 67W more than the GTX 680 at the wall is not insignificant.
For a change of perspective we shift over to OCCT, which is our standard pathological workload and almost entirely GPU-driven. Compared to Metro, the power consumption increase from the 7970 to the 7970GE isn’t as great, but it’s definitely still there. Power has increased by 19W at the wall, which is actually more than we would have expected given the fact that the two have the same PowerTune limit and the fact that PowerTune should be heavily throttling both cards. Consequently this means that the 7970GE creates an even wider gap between the GTX 680 and AMD’s top card, with the 7970GE pulling 43W more at the wall.
Moving on to temperatures, we don’t see a major change here. Identical hardware begets identical idle temperatures, which for the 7970GE means a cool 34C. Though the GTX 680 is a smidge cooler at 32C.
Since we’ve already seen that GPU power consumption has increased under Metro, we would expect temperatures to also increase under Metro and that’s exactly what’s happened. And actually, temperatures have increased by quite a lot, from 74C on the 7970 to 81C on the 7970GE. Since both 7970 cards share the same cooler, the 7970GE has to work harder to dissipate that extra power the card consumes, and even then temperatures will still increase some. 81C is still rather typical for a high end card, but it means there’s less thermal headroom to play with when overclocking when compared to the 7970. Furthermore it means the 7970GE is now warmer than the GTX 680.
Thanks to PowerTune throttling the 7970GE doesn’t increase in temperature by nearly as much under OCCT as it does Metro, but we still see a 4C rise, pushing the 7970GE to 83C. Again this is rather normal for a high-end card, but it’s a sign of what AMD had to sacrifice to reach this level of gaming performance.
Last but not least we have our look at noise. Again with the same hardware we see no shift in idle noise, with the 7970GE registering at a quiet 40.2dBA.
Unfortunately for AMD, this is where the 7970GE starts to come off of the rails. It’s not just power consumption and temperatures that have increased for the 7970GE, but load noise too. And it’s by quite a lot. 61.5dBA is without question loud for a video card. In fact the only card in our GPU 12 database that’s louder is the Radeon HD 6990, a dual-GPU card that was notoriously loud. The fact of the matter is that the 7970GE is significantly louder than any other card in our benchmark suite, and in all likelihood the only card that could surpass it would be the GTX 480. As a result the 7970GE isn’t only loud but it’s in a category of its own, exceeding the GTX 680 by nearly 10dBA! Even the vanilla 7970 is 6.3dBA quieter.
Does OCCT end up looking any better? Unfortunately the answer is no. At 63.2dBA it’s still the loudest single-GPU card in our benchmark suite by nearly 3dBA, and far, far louder than either the GTX 680 or the 7970. We’re looking at a 10.7dBA gap between the 7970GE and the GTX 680, and a still sizable 5.9dBA gap between the 7970GE and 7970.
From these results it’s clear where AMD has had to make sacrifices to achieve performance that could rival the GTX 680. By using the same card and cooler and at the same time letting power consumption increase to feed that speed, they have boxed themselves into a very ugly situation where the only solution is to run their cooler fast and to run it loud. Maybe, maybe with a better cooler they could have kept noise levels similar to the 7970 (which would have meant it would still be louder than the GTX 680), but that’s not what we’re looking at.
The 7970GE is without question the loudest single-GPU video card we have seen in quite some time, and that’s nothing for AMD to be proud of. Everyone’s limit for noise differs, but when we’re talking about single-GPU cards exceeding 60dB in Metro we have to seriously ponder whether it’s something many gamers would be willing to put up with.
OC: Power, Temperature, & Noise
Our final task is our look at the 7970GE’s overclocking capabilities. As the 7970GE is based on the existing 7970 we aren’t expecting any significant changes, however it’s reasonable to expect that general manufacturing process improvements over the last 6 months will have pushed yields and tolerances a little higher, giving us just a bit more headroom.
At the same time the presence of the boost clock and its associated voltage is going to change overclocking as well. The higher voltage should lend itself to higher overclocks, meanwhile validating overclocks is also going to be a bit harder as now we need to make sure neither the overclocked base clock/voltage combination or the overclocked boost clock/voltage combination is unstable, similar to the extra effort needed to overclock the GTX 680 series.
|Radeon HD 7970 Series Overclocking|
|Ref 7970GE||Ref 7970||XFX 7970 BEDD|
|Shipping Core Clock||1000MHz||925MHz||1000MHz|
|Shipping Max Boost Clock||1050MHz||N/A||N/A|
|Shipping Memory Clock||6GHz||5.5GHz||5.7GHz|
|Shipping Max Voltage||1.218v||1.175v||1.175v|
|Overclock Core Clock||1150MHz||1100MHz||1125MHz|
|Overclock Max Boost Clock||1200MHz||N/A||N/A|
|Overclock Memory Clock||6.4GHz||6.3GHz||6.3GHz|
|Overclock Max Boost Voltage||1.218v||1.175v||1.175v|
After going through the full validation process we were able to hit an overclock of +150MHz, which pushed our base clock from 1000MHz to 1150MHz, and our boost clock from 1050MHz to 1200MHz. Depending on how you want to count this overclock amidst the presence of the boost clock this is either 25MHz better than our best 7970 card, or 75MHz better. In either case our 7970GE definitely overclocks better than our earlier 7970 cards but not significantly so, which is in-line with our expectations.
As with any overclocking effort based on a single sample our overclocking results are not going to be representative of what every card can do, but they are reasonable. With AMD now binning chips for the 7970GE we’d expect to see some stratification among the 7970 family such that high overclocking chips that would previously show up in 7970 cards will now show up in 7970GE cards instead. For penny-pinching overclockers this is not good news, but for more hardcore overclockers this is nothing new as AMD’s partners have been doing something similar with their factory overclocked cards for some time now.
Meanwhile our memory overclock isn’t significantly different from what we could pull off with the reference 7970. The limitation is the memory bus or Tahiti’s memory controller, neither of which has changed. After around 6.4GHz errors start catching up and performance gains become performance losses.
Moving on to our performance charts, we’re going to once again start with power, temperature, and noise, before moving on to gaming performance. We’ll be testing our 7970 cards with the PowerTune limit set to +20% in order to avoid any real possibility of being performance limited by PowerTune.
With the 7970GE’s already high load power, overclocking and raising the PowerTune limits isn’t doing it any favors when it comes to overclocking. On the contrary to being a free overclock power consumption now exceeds even the GTX 690 in all situations and power consumption is almost certainly in excess of 300W at the card level. As we’ll see in our gaming performance section we’re definitely getting more performance out of the 7970GE, but we’re paying for it with power.
With a rise in power consumption comes a rise in temperatures to a varying degree. At 83C under Metro the 7970GE has gotten warmer, but not significantly so. The same cannot be said for OCCT. At 89C we’re approaching the reasonable limits for this card and cooler.
The 7970GE was already loud at stock and overclocking it doesn’t help. Under Metro noise is now at 63.8dBA, and under OCCT it’s tied with the 6990 for noise at 66dBA. Even if you’re forgiving of noise, this is reaching the point where it’s going to be difficult to ignore. Serious 7970GE overclockers will want to seek other cards and/or aftermarket coolers.
OC: Gaming Performance
Moving on to our gaming performance section, let’s see what our 15% overclock can do for gaming performance. Assuming the 7970GE’s performance continues to scale well with clockspeeds, we should see around a 10-12% real world performance gain.
Starting as always with Crysis, what was already an incredible lead for the 7970GE becomes increasingly absurd. The 7970GE OC sees a further 8% performance improvement here, pushing it past 60fps at 1920 and past 40fps at 2560. Against the stock GTX 680 it’s now leading by 35% on average framerates.
Minimum framerates on the other hand haven’t changed by nearly as much. Given what we know about PT Boost and Crysis’s love of memory bandwidth, it’s reasonable to assume we’re being held back by memory bandwidth limits here.
Moving on to Metro, the 7970GE once again gains 8-9% due to our overclock. The earlier 56W jump in power consumption at the wall makes it an expensive gain, but as it stands this is by far the fastest we’ve seen a single-GPU card perform at Metro.
With Batman we once more see gains of around 9% from our overclock. Even at 2560 our framerates are now in excess of 70fps.
With Portal 2 being one of the 7970GE’s biggest defecits relative to the GTX 680, a solid overclock can help to close the gap but it’s not enough to eliminate it. Still, it’s enough to push the average framerate over 60fps at 2560 with SSAA. The overall scaling from the overclock also looks very good here, with the 7970GE OC picking up a larger than normal 11-12%.
Finally we have AMD’s other notable weak spot, Battlefield 3. Much like with Portal 2 overclocking can’t eliminate the GTX 680’s lead, but it can significantly cut into it. At 2560 and at 1920 with MSAA this is enough to push past 60fps, making for a solid 10% performance gain from overclocking.
Bringing things to a close, before writing up this article I spent some time going through our archives to take a look at past GPU reviews. While AMD has routinely retaken the performance crown for a time by beating NVIDIA in releasing next-generation GPUs first – such was the case with the Radeon HD 5870 and Radeon HD 7970 – the typical pattern is for AMD’s flagship single-GPU card to trail NVIDIA’s flagship once NVIDIA has caught up. In a generational matchup AMD has not been able to beat or tie NVIDIA for the highest performing single-GPU card a very long time. And as it turns out the last time that happened was six years ago, with the Radeon X1950 XTX in 2006.
Six years is a long time to wait, but patience, perseverance, and more than a few snub moves against NVIDIA have paid off for AMD. For the first time in 6 years we can say that AMD is truly competitive for the single-GPU performance crown. The Radeon HD 7970 GHz Edition isn’t quite fast enough to outright win, but it is unquestionably fast enough to tie the GeForce GTX 680 as the fastest single-GPU video card in the world today. With that said there’s a lot of data to go through, so let’s dive in.
As far as pure gaming performance goes the 7970GE and the GTX 680 are tied in our benchmarks at the top single monitor resolution of 2560x1600. The 7970GE scores some impressive wins in Crysis and DiRT 3, while NVIDIA manages to hold on to their substantial leads in Battlefield 3 and Portal 2. Elsewhere we see the 7970GE win at some games while the GTX 680 wins at others, and only very rarely do the two cards actually tie. Ultimately this is very much a repeat of what we saw with the GTX 670 versus the 7970, and the 6970 versus the GTX 570, which is to say that the 7970GE and GTX 680 are tied on average but are anything but equal.
Our advice then for prospective buyers is to first look at benchmarks for the games they intend to play. If you’re going to be focused on only a couple of games for the near future then there’s a very good chance one card or the other is going to be the best fit. Otherwise for gamers facing a wide selection of games or looking at future games where their performance is unknown, then the 7970GE and GTX 680 are in fact tied, and from a performance perspective you couldn’t go wrong with either one.
As an addendum to that however, while the 7970GE and GTX 680 are tied at 2560x1600 and other single-monitor resolutions the same cannot be said for multi-monitor configurations. The 7970GE and GTX 680 still trade blows on a game-by-game basis with Eyefinity/NVIDIA Surround, but there’s a clear 6% advantage on average for the 7970GE. Furthermore the 7970GE has 3GB of VRAM versus 2GB for the GTX 680, which makes the 7970GE all the better suited for multi-monitor gaming in the future. AMD may be tied for single-monitor gaming, but they have a clear winner on their hands for multi-monitor gaming.
With that said, AMD has made a great sacrifice to get to this point, and it’s one that’s going to directly impact most users. AMD has had to push the 7970GE harder than ever to catch up to the GTX 680, and as a result the 7970GE’s power consumption and noise levels are significantly higher than the GTX 680’s. It’s unfortunate for AMD that NVIDIA managed to tie AMD’s best gaming performance with a 104-series part, allowing them to reap the benefits of lower power consumption and less noise in the process. Simply put, the 7970GE is unquestionably hotter and uncomfortably louder than the GTX 680 for what amounts to the same performance. If power and noise are not a concern then this is not a problem, but for many buyers they're going to be unhappy with the 7970GE. It’s just too loud.
Of course this isn’t the first time we’ve had a hot and loud card on our hands – historically it happens to NVIDIA a lot, but when NVIDIA gets hot and loud they bring the performance necessary to match it. Such was the case with the GTX 480, a notably loud card that also had a 15% performance advantage on AMD’s flagship. AMD has no such performance advantage here, and that makes the 7970GE’s power consumption and noise much harder to justify even with a “performance at any cost” philosophy.
The end result is that while AMD has tied NVIDIA for the single-GPU performance crown with the Radeon HD 7970 GHz Edition, the GeForce GTX 680 is still the more desirable gaming card. There are a million exceptions to this statement of course (and it goes both ways), but as we said before, these cards may be tied but they're anything but equal.
Noise issues aside, we’re finally seeing something that we haven’t seen for a very long time: bona fide, cut throat, brutal competition in the high-end video card segment for the fastest single-GPU video card. To call it refreshing is an understatement; it’s nothing short of fantastic. For the first time in six years AMD is truly performance competitive with NVIDIA at the high-end and we couldn't be happier.
Welcome back to the fight AMD; we’ve missed your presence.