The AMD Radeon R9 Fury Review, Feat. Sapphire & ASUS
by Ryan Smith on July 10, 2015 9:00 AM ESTA bit over two weeks ago AMD launched their new flagship video card, the Radeon R9 Fury X. Based on the company’s new Fiji GPU, the R9 Fury X brought with it significant performance improvements to AMD’s lineup, with AMD’s massive Fiji greatly increasing the card’s shading resources. Meanwhile Fiji also marked the introduction of High Bandwidth Memory (HBM) in to consumer products, giving the R9 Fury X a significant leg up in memory bandwidth. Overall AMD put together a very impressive card, however at $649 it fell just short of the GeForce GTX 980 Ti AMD needed it to beat.
Meanwhile alongside the announcement of the R9 Fury X, AMD announced that there would be three other Fiji-based cards. These include the R9 Fury, the R9 Nano, and a yet-to-be-named dual-GPU Fiji card. The first of these remaining cards to launch would be the R9 Fury, the obligatory lower-tier sibling to AMD’s flagship R9 Fury X. Today we will be taking a look at the first of those remaining cards, the R9 Fury, which launches next week.
While R9 Fury X remains the fastest Fiji card – and by virtue of being introduced first, the groundbreaking card – the impending launch of the R9 Fury brings with it a whole slew of changes that make it an interesting card in its own right, and a very different take on a Fiji product altogether. From a performance standpoint it is a lower performing card, featuring a cut-down Fiji GPU, but at the same time it is $100 cheaper than the R9 Fury X. Meanwhile in terms of construction, unlike the R9 Fury X, which is only available in its reference closed loop liquid cooling design, the R9 Fury is available as semi-custom and fully-custom cards from AMD’s board partners, built using traditional air coolers, making this the first air cooled Fiji card. As a result the R9 Fury at times ends up being a very different take on Fiji, for all of the benefits and drawbacks that comes with.
AMD GPU Specification Comparison | ||||||
AMD Radeon R9 Fury X | AMD Radeon R9 Fury | AMD Radeon R9 290X | AMD Radeon R9 290 | |||
Stream Processors | 4096 | 3584 | 2816 | 2560 | ||
Texture Units | 256 | 224 | 176 | 160 | ||
ROPs | 64 | 64 | 64 | 64 | ||
Boost Clock | 1050MHz | 1000MHz | 1000MHz | 947MHz | ||
Memory Clock | 1Gbps HBM | 1Gbps HBM | 5Gbps GDDR5 | 5Gbps GDDR5 | ||
Memory Bus Width | 4096-bit | 4096-bit | 512-bit | 512-bit | ||
VRAM | 4GB | 4GB | 4GB | 4GB | ||
FP64 | 1/16 | 1/16 | 1/8 | 1/8 | ||
TrueAudio | Y | Y | Y | Y | ||
Transistor Count | 8.9B | 8.9B | 6.2B | 6.2B | ||
Typical Board Power | 275W | 275W | 250W | 250W | ||
Manufacturing Process | TSMC 28nm | TSMC 28nm | TSMC 28nm | TSMC 28nm | ||
Architecture | GCN 1.2 | GCN 1.2 | GCN 1.1 | GCN 1.1 | ||
GPU | Fiji | Fiji | Hawaii | Hawaii | ||
Launch Date | 06/24/15 | 07/14/15 | 10/24/13 | 11/05/13 | ||
Launch Price | $649 | $549 | $549 | $399 |
Starting things off, let’s take a look at the specifications of the R9 Fury. As we mentioned in our R9 Fury X review, we have known since the initial R9 Fury series launch that the R9 Fury utilizes a cut-down Fiji GPU, and we can now reveal just how it has been cut down. As is usually the case for these second-tier cards, the R9 Fury features both a GPU with some functional units disabled and a slightly reduced clockspeed, allowing AMD to recover partially defective GPUs while easing up on the clockspeed requirements.
The Fiji GPU in the R9 Fury ends up having 56 of 64 CUs enabled, which brings down the total stream processor count from 4,096 to 3,584. This in turn ends up being the full extent of the R9 Fury’s disabled functional units, as AMD has not touched the front-end or back-end, meaning the number of geometry units and the number of ROPs remained unchanged.
Also unchanged is the memory subsystem. All Fiji-based cards, including the R9 Fury, will be shipping with a fully enabled memory subsystem, meaning we’re looking at 4GB of HBM attached to the GPU over a 4096-bit memory bus. With Fiji topping out at just 4GB of memory in the first place – one of the drawbacks faced by the $650 R9 Fury X – cutting back on memory here to a smaller capacity is not a real option for AMD, so every Fiji card will come with that much memory.
As for clockspeeds, R9 Fury takes a slight trim on the GPU clockspeed. The reference clockspeed for the R9 Fury is a flat 1000MHz, a 5% reduction from the R9 Fury X. On the other hand the memory clock remains unchanged at 500MHz DDR, for an effective memory rate of 1Gbps/pin.
All told then, on paper the performance difference between the R9 Fury and R9 Fury X will stand to be between 0% and 17%; that is, the R9 Fury will be up to 17% slower than the R9 Fury X. In the best case scenario for the R9 Fury of a memory bandwidth bottleneck, it has the same 512GB/sec of memory bandwidth as the R9 Fury X. At the other end of the spectrum, in a shader-bound scenario, the combination of the reduction in shader hardware and clockspeeds is where the R9 Fury will be hit the hardest, as its total FP32 throughput drops from 8.6 TFLOPs to 7.17 TFLOPs. Finally in the middle, workloads that are front-end or back-end bound will see a much smaller drop since those units haven’t been cut-down at all, leading to just a 5% performance drop. As for the real world performance drop, as we’ll see it’s around 7%.
Power consumption on the other hand is going to be fairly similar to the R9 Fury X. AMD’s official Typical Board Power (TBP) for the R9 Fury is 275W, the same as its older sibling. Comparing the two products, the R9 Fury sees some improvement from the disabled CUs, however as a second-tier part it uses lower quality chips overall. Meanwhile the use of air cooling means that operating temperatures are higher than the R9 Fury X’s cool 65C, and as a result power loss from leakage is higher as well. At the end of the day this means that the R9 Fury is going to lose some power efficiency compared to the R9 Fury X, as any reduction in power consumption is going to be met with a larger decrease in performance.
Moving on, let’s talk about the cards themselves. With the R9 Fury X AMD has restricted vendors to selling the reference card, and we have been told it will be staying this way, just as it was for the R9 295X2. On the other hand for R9 Fury AMD has not even put together a complete reference design, leaving the final cards up to their partners. As a result next week’s launch will be a “virtual” launch, with all cards being semi or fully-custom.
Out of the gate the only partners launching cards are Sapphire and Asus, AMD’s closest and largest partners respectively. Sapphire will be releasing stock and overclocked SKUs based on a semi-custom design that couples the AMD reference PCB with Sapphire’s Tri-X cooler. Asus on the other hand has gone fully-custom right out of the gate, pairing up a new custom PCB with one of their DirectCU III coolers. Cards from additional partners will eventually hit the market, but not until later in the quarter.
The R9 Fury will be launching with an MSRP of $549, $100 below the R9 Fury X. This price puts the R9 Fury up against much different competition than its older sibling; instead of going up against NVIDIA’s GeForce GTX 980 Ti, the closest competition will be the older GeForce GTX 980. The official MSRP on that card is $499, so the R9 Fury is more expensive, but in turn AMD is promising better performance than the GTX 980. Otherwise NVIDIA’s partners serve to fill that $50 gap with their higher-end factory overclocked GTX 980 cards.
Finally, today’s reviews of the R9 Fury are coming slightly ahead of the launch of the card itself. As previously announced, the card goes on sale on Tuesday the 14th, however the embargo on the reviews is being lifted today. AMD has not officially commented on the launch supply, but once cards do go on sale, we’re expecting a repeat of the R9 Fury X launch, with limited quantities that will sell out within a day. After that, it seems likely that R9 Fury cards will remain in short supply for the time being, also similar to the R9 Fury X. R9 Fury X cards have come back in stock several times since the launch, but have sold out within an hour or so, and there’s currently no reason to expect anything different for R9 Fury cards.
Summer 2015 GPU Pricing Comparison | |||||
AMD | Price | NVIDIA | |||
Radeon R9 Fury X | $649 | GeForce GTX 980 Ti | |||
Radeon R9 Fury | $549 | ||||
$499 | GeForce GTX 980 | ||||
Radeon R9 390X | $429 | ||||
Radeon R9 290X Radeon R9 390 |
$329 | GeForce GTX 970 | |||
Radeon R9 290 | $250 | ||||
Radeon R9 380 | $200 | GeForce GTX 960 | |||
Radeon R7 370 Radeon R9 270 |
$150 | ||||
$130 | GeForce GTX 750 Ti | ||||
Radeon R7 360 | $110 |
288 Comments
View All Comments
Socius - Saturday, July 11, 2015 - link
Yes...comparing an overclocked 3rd party pimped model against a non-OC'd reference design card that is $100 less in price and then saying it's a tough choice between the 2 as the R9 Fury is faster than the GTX 980...lol...totally balanced perspective there bro.Ryan Smith - Saturday, July 11, 2015 - link
Whenever a pure reference card isn't available, like is the case for R9 Fury, we always test a card at reference clockspeeds in one form or another. For example in this article we have the ASUS STRIX, which ships at reference clockspeeds. No comparisons are made between the factory overclocked Sapphire card and the GTX 980 (at the most we'll say something along the lines of "both R9 Fury cards", where something that is true for the ASUS is true for the Sapphire as well).And if you ever feel like we aren't being consistent or fair on that matter, please let us know.
Socius - Saturday, July 11, 2015 - link
I just want to understand something clearly. Are you saying that when you are talking about performance/value, you ignore the fact that one card can OC 4%, and the other can OC 30% as you don't believe it's relevant? Let me pose the question another way. If you had a friend who was looking to spend $500~ on a card and was leaning between the R9 Fury and the GTX 980. Knowing that the GTX 980 will give him better performance once OC'd, at $450, would you even consider telling him to get the R9 Fury at $550?My concern here is that you're not giving a good representation of real world performance gamers will get. As a result, people get misled into thinking spending the extra money on the R9 Fury is actually going to net them higher frame rate...not realizing they could get better performance, for even less money, if someone decided to actually look at overclocking potential...
Now if you weren't interested in overclocking results in general, I'd say fine. I disagree, but it's your choice. But then you do show overclocking results with the R9 Fury. I'm finding it really hard to understand what your intent is with these articles, if not to educate people and help them make an actual informed decision when making their next purchase.
As I mentioned in your previous article on the Fury X...you seem to have a soft spot for AMD. And I'm not exactly sure why. I will admit that I'm currently a big Nvidia fan. Only because of the features and performance I get. If the Fury X had come out, and could OC like the 980ti and had 8gb HBM memory, I'd have become an AMD fan. I'm a fan of whoever has the best technology at any given moment. And if I were looking to make a decision on my next card purchase, your article here would give a false impression of what I would get if I spent $100 more on an R9 fury, than on a GTX 980...
jardows2 - Saturday, July 11, 2015 - link
Apparently, OC is your thing. I get it. There are plenty of OC sites that are just for you. For some of us, we really don't want to see how much we can shorten the life of something we pay good money for, when the factory performance does what we need. I, for one, am more interested in how something will perform without me potentially damaging my computer, and I appreciate the way that AT does their benchmarks.Socius - Saturday, July 11, 2015 - link
I think the fact that you believe overclocking will "damage your computer" or in any meaningful way shorten the lifespan of the product, is all the more reason to talk about overclocking. I'd be more than welcome to share a little info.Generally speaking, what kills the product is heat (minus high current degradation that was a bigger problem on older fab processes) So let's say you have a GPU that runs 60 degrees Celsius under load, at 1000MHz with a 50% fan speed profile. Now let's imagine 2 scenarios:
1) You underclock the GPU to 900MHz and set a 30% fan profile to make your system more quiet. Under load, your GPU now hits 70 degrees Celsius.
2) You overclock the GPU to 1100MHz and set a 75% fan profile for more performance at the cost of extra sound. Under load, your GPU now hits 58 degrees Celsius due to increased fan speed.
Which one of these devices would you think is likely to last the longest? If you said the Overclocked one, you'd be correct. In fact...the overclocked one is likely to last even longer than the stock 1000MHz at the 50% fan speed profile, because despite using more power and giving more performance, the fan is working harder to keep it cooler, thus reducing the stress on the components.
Now. Let's talk about why that card was clocked at 1000MHz to start! When a chip is designed, the exact clock speed is an unknown. Not just between designs...but between individual wafers and dies cut out from those wafers. As an example, I had an i7 3770k that would use 1.55v to hit 4.7GHz. I now have one that uses 1.45v to hit 5.2GHz and 1.38v to hit 5GHz. Why am I telling you this? Well...because it's important! When designing a product like a CPU or GPU, and setting a base clock, you have to account for a few things:
1) How much power do I want to feed it?
2) How much of the heat generated by that power can I dissipate with my fan design?
3) How loud do I want my fan to be?
4) What's the highest clock rate my lowest end wafer can hit while remaining stable and at an acceptable voltage requirement?
So here's the fun part. So while chips themselves can vary greatly, there are tons of precautions added when dealing with stock speeds. Just for a point of reference...a single GTX Titan X is guaranteed to overclock 1400MHz-1550MHz with proper settings, if you put the fan at 100% full blast. That's a 30%-44% overclock! So why wouldn't Nvidia do that? Well it's a few things.
1) Noise! Super important here. Your clockspeed is determined by your ability to cool it. And if you're cooling by means of a fan, the faster that fan, the more noise, the more complaints by consumers.
2) Power/Heat variability. Since each chip is different, as you go into the higher ranges, each will require a different amount of power in order to be stable at that frequency. If you're curious, you can see what's called an ASIC quality for your GPU using a program like GPU-Z. This number will tell you roughly how good of a chip you have, in terms of how much of a clock it can achieve with how much power. The higher the % of your ASIC quality, the better overclocking potential you have on air because it'll require less power, and therefore create less heat to do it!
3) Overclocking potential. This is actually important in Marketing. And it's something AMD and Nvidia are both pretty bad at, actually. But AMD's bit a bit worse, to their own detriment. In their R9 Fury and Fury X press release performance numbers they set expectations for their cards to completely outperform the 980ti using best case scenarios and hand picked settings. And they also said it overclocks like a beast. Now...here's why that's bad. Customers like to feel they're getting more than what they pay for. That's why companies like BMW always list very modest 0-60 times for their cars. When I say modest, I mean they set the 0-60 times to show they're worse than what the car is actually capable of. That's why every car review program you see will show 0-60 times being 0.2 to 0.5 seconds faster than what BMW has actually listed. This works because you're sold on a great product, only to find it's even greater than that.
Got off track a bit there. I apologize. Back to AMD and the Fiji lineup and why this long post was necessary. When AMD announced the Fury X being an all in one cooler design, I instantly knew what was up. The chip wasn't able to hold up to the limitations we talked about above (power requirement/heat/fan noise/stability). But they needed to put out a stock clock that would allow the card to be competitive with Nvidia, but they also didn't want it to sound like a hair dryer. That's why they opted for the all in one cooler design. Otherwise, a chip that big on an air cooled design would likely have been clocked around the 850-900MHz range that the original GTX Titan had. But they wanted the extra performance, which created extra heat and required more power, and used a better cooler design to be able to accomplish that across the board with all their chips. That's great, right? Well...yes and no. I'll explain.
Essentially the Fiji lineup is "factory overclocked" by AMD. This is the same as putting a turbo on a car engine. And as any car enthusiast will tell you, a 2L engine with a turbo on it may be able to produce 330 horsepower, which could otherwise take a naturally aspirated 4L V8 to accomplish. But then you're limited for increased horsepower even further. Sure you could put a bigger supercharger on that 2L engine. But it already had a boosted performance for that size engine. So your gains will be minimal. But with that naturally aspirated engine...you can drop a supercharger on it and realize massive gains. This is very much the same as what's happening with the Fury X, for example.
And this is why I believe it's incredibly important to point this out. I built a system for my friend recently with a GTX 970. I overclocked that to 1550MHz on first try, without even maxing out the voltage. That was a $300 model card. And even that would challenge the performance of the R9 Fury if you don't plan on overclocking (not that you could, anyway, with that 4% overclock limit).
So...yes, I do think Overclocking needs to be talked about more, as it's become far easier and safer to do than in the past. Even if you don't plan to do extreme overclocking, you can keep all your fan speed automated profiles the same, don't touch voltage, and just increase power limit and a slight increase in the gpu clock for simple free performance. It's something I could teach my mom to do. So I hope it's something that is done by more and more people, as there's really no reason not to do it.
And that's why I think informing people of these differences with regard to overclocking will help people save money, and get more performance. And not doing just keeps people in the dark, and does them a great disservice. Why keep your audience oblivious and allow them to remain ignorant to these things when you can take some time to help them? Overclocking today is far different from the overclocking of a few years ago. And everybody should give it a try.
mdriftmeyer - Sunday, July 12, 2015 - link
Ryan is biased in these comparisons. That's a fact. Your knowledge silencing the person defending Ryan's unbalanced comparisons is a fact.jardows2 - Monday, July 13, 2015 - link
What a hoot! He didn't silence me. I just have better things to do on the weekend than check all the Internet forums I may have posted to!jardows2 - Monday, July 13, 2015 - link
@SociusThat was a good post explaining your interest in OC. From my perspective, I come from the days when OC'ing would void your warranty and would shorten the life span of your system. I also come from a more business and channel oriented background, meaning that we stay with "officially supported" configurations. That background stays with you. Even today, when building my personal computer or a web browsing computer for a family member, I do everything I can to stay with the QVL of the motherboard.
I am more concerned with out of box experience, and almost always skip over the OC results of benchmarks whenever they are presented. If a device can not perform properly OOB, then it was not properly configured from the start, which does not give me the best impression about the part, regardless of the potential of individual tweaks.
In the end, you are looking for OC results, I am only concerned with OOB experience. Different target audience, both with valid concerns. I just don't think it is worth bashing a reviewer's method that focuses on one experience over the other.
Mugur - Saturday, July 11, 2015 - link
Good review as always, Ryan. I wouldn't throw away Asus's approach. Nice power efficiency gains. Funny how overclocking for a few fps more heated the discussion... All in all, I think AMD is back in business with Fury non-X. Waiting for 3xx reviews, hopefully for the whole R9 line, with 2, 4 and 8 GB.3ogdy - Saturday, July 11, 2015 - link
What a disappointment again. Man...the Fury cards really aren't worth the hassle at all, are they? It's sad to see this, especially coming from an FX-8350 & 2xHD6950s owner. So the 980Ti beats the custom cards (we knew Fury X was quite at its limits from the beginning, but still - despite all the improvements, custom cards sometimes perform even worse than the stock one. The 980Ti really beats the Fury X most of the time.What is that, nVidia's blower is only 8dB louder than the highest end of the fans used by arguably the two best board partners AMD has? Wow! This is where I realize nVidia really must have done an amazing job with the 980Ti and Maxwell in general.
HBM this HBM that...this card is beaten at 4K by the GTX980Ti and gameplay seems to be smoother on the nVidia card too. What the hell? Where are the reasons to buy any of the Fury cards?