Conclusion
Conclusion
AMD’s R9 Fury X was launched amidst a huge amount of optimism but, despite being a good graphics card, it ultimately failed to distinguish itself from the competition. In the eyes of many, it is too expensive for what’s being offered performance-wise and AMD’s feature set tends to lag behind in many ways. While the R9 Fury exhibits the same feature-bound limitations as its sibling, it is infinitely more appealing due to three simple factors: price, performance relative to competing solutions and its lack of a clunky water cooling solution.
Let’s start with the 500 pound gorilla in the room: the Fiji architecture’s abysmal performance in some key titles. The R9 Fury X could have been an outstanding GPU had it not been for sub-standard framerates in games like Battlefield 4, Grand Theft Auto V and Shadow of Mordor. Now before you go on a tangent and blame a lack of driver maturity for any missteps, consider this: the Fiji’s design is essentially an upscaled and very slightly updated version of a core that’s been around for more than two years now. If AMD still hasn’t been able to squeeze optimal performance out of such an old architecture, there are some serious problems that need to be addressed. Drivers should be a solution, not an excuse.
With all of that being said nothing has changed this time around (particularly at 1440P) and the R9 Fury finds itself lagging behind in those same situations, though sometimes not as drastically. Luckily, its substantially lower price makes any disparity between the new Radeon and GeForce lineups all that much more palatable. Indeed AMD’s $549 price for this thing is absolutely perfect, laser targeting a yawning $150 gap between the GTX 980 Ti and GTX 980.
With the GTX 980 and GTX 980 Ti safely bracketing it, the R9 Fury represents an excellent solution for anyone that wants a bit more power than the GTX 980 or R9 390X but can’t justify the $649 entry cost of the R9 Fury X or GTX 980 Ti. At 1440P it performs closer to the GTX 980, though still handily beating the GM204-based NVIDIA card.
Like its larger sibling the Fury’s stance improves when switching over to 4K where it is within spitting distance of the 980 Ti and extends an already-impressive lead over the GTX 980. Granted, there are plenty of pre-overclocked GTX 980’s on the market that can somewhat bridge the gap but they’re typically just as expensive as a stock-clocked R9 Fury.
In many ways these numbers will put more pressure on the R9 Fury X than they will on anything within NVIDIA’s current stable. This card comes within only a few percentage points of its bigger brother and with a bit of overclocking (granted, there isn’t much left in the frequency tank) the situation becomes even tighter. Perhaps even more importantly the R9 Fury does without that horribly clunky closed loop liquid cooler, effectively eliminating all of the installation, longevity and pump noise headaches it brings to the table.
Many wondered why the R9 390X was priced so high relative to outgoing R9 290X cards and now you see the reason. I’ll take the lower power consumption of Fiji over Hawaii any day but the 390X’s price / performance ratio relative to the Fury may cause some folks to jump ship in order to save over a hundred bucks. At 1440P it loses by a scant 10%, at times nearly matching the Fury’s framerates.
Sapphire’s R9 Fury Tri-X is an interesting card to say the least. Since the Fiji Pro core will always be paired up with a board partner-designed cooling solution we’ll likely see plenty of options but Sapphire’s Tri-X cooler cooler proved to be a willing participant in the review. There’s no escaping the fact that AMD’s new architecture runs stupidly hot but this heatsink shrugged the extreme thermal load aside while still delivering excellent acoustical results and some room for overclocking. This card’s titanic size and raw thermal dissipation mass also makes me wonder how much will need to be cut from the Fury Nano in order for it to achieve reasonable temperatures in such a compact form factor. But that’s another conversation for another time…
The real question is whether or not Sapphire’s R9 Fury Tri-X OC is worth a premium over the Tri-X equipped stock-clocked version. In terms of real-world framerates its paltry 4% overclock equates about 2% and at the very most 3% better performance. Considering NVIDIA’s board partners are achieving substantially higher thresholds on their OC SKUs, we have to wonder where this hesitation is coming from and just how restrictive this architecture really is to overclocking. One way or another, for these kind of numbers there’s just no reason to spend more money on Sapphire’s OC model. Stick with their perfectly capable reference-clocked board.
If there’s any card that will put a modicum of fear into NVIDIA, it will be the R9 Fury since it successfully attacks a price point that currently lacks a GeForce alternative. It may have been designed to alleviate some sales pressure from the R9 Fury X but one has to wonder whether or not it will render AMD’s flagship somewhat disposable in short order. This is a perfectly targeted graphics card that, when paired up with Sapphire’s awesome Tri-X cooler, combines excellent performance results and whisper quiet operation into a product that should be poised for huge success.