There is a very real maxwell advantage that is huge when you think about it.
Well most people on our forums know that Kepler is a write-off against 290X or 970. It's virtually impossible to recommend Kepler GPUs for anything anymore. As far as Maxwell having any future advantage over GCN, Fury clearly has a major advantage over 980 and it's evident once we look at performance in 4K. The potential of a card to run future games well can be seen in GPU demanding situations. Once we move from 1440P to 4K, 980
falls apart completely but the regular Fury is very close to a 980Ti/Fury X.
When we look at specific modern GPU demanding titles, 980 isn't even in the same league, often with its averages only as good as Fury's minimums:
Even PCPerspective agrees 980 is
no competition for the Fury at high resolution gaming:
"Based on the performance data we are seeing here today, where the AMD Radeon R9 Fury ranges from being 5-30% faster than the GTX 980, I think AMD has released a part that puts a lot of pressure on NVIDIA's GeForce team. It is more expensive, but for $50 you are getting all custom, retail coolers as well as improved performance in the vast majority of our testing." ~
PCPer
^ If PCPer is saying Fury smashes a 980 by up to 30% at high resolutions, well it doesn't get more definitive than that considering they tend to favour NV products. Once we consider 980 SLI vs. Fury CF, it's not even a discussion ==> Fury CF all the way.
Also, there is some serious cherry-picking by certain posters happening in this thread who are focusing in on the Asus Strix $580 but Sapphire Tri-X Fury is $549.
Also, I see how some are ignoring that the Sapphire Tri-X Fury is the quietest flagship card that's ever been tested by AT.
DX12 drivers are likely to benefit GCN more than 980 because it's AMD that has a larger DX11 overhead. Finally, it's guaranteed that AMD will continue supporting GCN for years to come but who is going to guarantee that NV will support Maxwell once Pascal launches?
And this is before we even touch how NV tried to block mobile dGPU overclocking, lied about Batman AK visuals to market the game (showed off all their demos at 60 fps), lied and showed no remorse for 970 fiasco.
We also don't know how well Fury cards really overclock until someone releases a voltage unlock. It's hard to imagine that Fury will continue to remain voltage locked forever.
There is little question imo that Fury is the better buy over the 980 for keeping over the next 3 years for 1440P because in almost all GPU demanding situations, 980 gets wrecked by Fury at 4K. If someone wants a good 1080/1200P card, 290X/390 or 970 are better bets anyway at far less $. The bigger obstacle for Fury is at that point might as well spend the extra $100 for a 980Ti and get 30%+ more performance in a 980Ti OC. That's the route I would go if I were to spend $550 today. 980 though continues to be an irrelevant card between the 290X/970 and 390. I do think that both the 980 and Fury need price cuts as both are priced way too high.
Relative to a 3-year-old $499 HD7970Ghz (aka 280X),
980 is only 55% faster and Fury is only 67% faster at 1440P. That's pretty pathetic for cards that cost $480-550 and well short of the 2X the performance increase over 3 years that we have had since the fall of 2009.
Unfortunately the market continues to be divided between R9 290/290X/390/970 and 980Ti with nothing in between really worth buying at current prices. I suppose for those who do not overclock, $1100 Fury CF can win a few customers but the card would have been a lot more attractive at $479-499. Looks like the same ground-breaking shake-up in the market that HD5850/6950 and R9 290 delivered is nowhere to be found here. However, we shouldn't forget that HD7950 launched at $449 but over time the card became a smash hit once it dropped towards $280-300 mark against much more expensive 670/680 cards. If in the next 6 months Fury can drop closer to $450 with rebates and discounts, then it could become a solid buy.