AMD Radeon Fury X Series | HBM, Small Form Factor And Water Cooling | June 16th

Am I the only one excited about the possible low to no noise coming out of the Fury X card? With the awesome low temps?
Coming from a GTX-680 will be a dream. Just need to wait for real reviews and see if I go Fury X or watercooled 980ti

Pretty exciting year with so many new toys :)
 
I doubt it. I think the best custom 980 Ti's (also water) will OC far more than Fury X. It's a matter of thermals, and the 980Ti uses less power at stock and has more room to go. That being said, this is $649. The top custom 980 Ti are $700+. AMD should win the $650 battle, but those willing to put more money in I think will find that Nvidia still has the performance crown. The question is if you want to drop extra cash, because Fury X has the best reference cooler and doesn't carry a price premium.

I mean, look at this:


There's no reason this won't beat a Fury X, and of course Titan X unless you custom modify the Titan X to water. Nvidia still holds the the top GPU performance crown if you are willing to pay for it.

Moving on, there's also a Hybrid Water 980Ti: if its cooler is as good as AMD's, should be a faster card:

http://www.newegg.com/Product/Produ...4487144&cm_re=gtx980ti-_-14-487-144-_-Product

But yeah, $120 premium. The Galaxy HOF card images I posted will probably have a similar extra cost if not higher.

But let's see how this does against $650 980 Ti's at max OC. Or even against typical aftermarket 980Ti's that are still under $700. I'm not even convinced the Fury X will out perform them at Max OC. Even a not so expensive aftermarket 980Ti has very high OC performance.



This card is only $30 more. Fury X is estimated to be 50-55% faster than a 290X, right? Let's take a look at the $680 980Ti:

55% faster / 72% OC

57% faster / 71% faster OC


Unfortunately that's all they tested with OC, but even with its "stock OC" it is right around that 55% faster than 290X (Fury X) speed:

60% faster


65% faster


36% faster


58% faster


Source: http://hexus.net/tech/reviews/graphics/83819-evga-geforce-gtx-980-ti-superclocked-acx-20/

Who'd spend $650 but not $680? It will be interesting to see how Fury X does max OC against cards like this - because this card OC's well and asks for a reasonable minor premium. Fury X is hybrid water so it should OC faster in a equal playing field, but has less thermal room. It'll be interesting to see who wins the $650 to <$700 battle. Although AMD surely has the noise battle won.

You're confusing the 290/390x for the fury. AMD released benches put the fury x ahead of the titan, and AMD is saying the fury x is built for overclocking. So you're pretty much just wrong here. Ocd fury x will beat a 980 ti, even if it's under water.

Edit: I reread and I get what your saying about estimates for fury x over 290x, but my point still stands that the fury x is beating a titan in witcher 3, bf4, etc. Benches are earlier in the thread.
 
You're confusing the 290/390x for the fury. AMD released benches put the fury x ahead of the titan, and AMD is saying the fury x is built for overclocking. So you're pretty much just wrong here. Ocd fury x will beat a 980 ti, even if it's under water.

He's not confusing them, he's saying these custom 980Ti's with crazy power delivery and water-cooling should theoretically beat the Fury X, which is probably true.

The problem is those custom 980Ti's are $100-200+ more expensive than the Fury X, so it's already entering a completely different pricing tier.
 
He's not confusing them, he's saying these custom 980Ti's with crazy power delivery and water-cooling should theoretically beat the Fury X, which is probably true.

The problem is those custom 980Ti's are $100-200+ more expensive than the Fury X, so it's already entering a completely different pricing tier.

Yeah I reread and edited my post. The benches AMD released suggest otherwise, and AMD is saying the fury x is a great overclocker. So I think even ocd vs ocd AMD will beat the 980 ti. A titan x under water might beat a fury x when both are max ocd, but then you're talking $1200+ vs $650.
 
amd-radeon-fury-x-4k-qcq6i.png


amd-radeon-fury-overcw7ren.png

(With 100MHz OC on core)

amd-radeon-fury-x-spl99p2a.jpg


http://www.pcworld.com/article/2937...x-tech-specs-and-design-details-revealed.html

So yeah, it doesn't seem like an ocd 980 ti will beat a fury x that's also ocd.
 
Yeah I reread and edited my post. The benches AMD released suggest otherwise, and AMD is saying the fury x is a great overclocker. So I think even ocd vs ocd AMD will beat the 980 ti. A titan x under water might beat a fury x when both are max ocd, but then you're talking $1200+ vs $650.

Yeah I'm thinking Nvidia always going cheap with their reference coolers is actually going to be a problem really soon now. I hope this actually motivates Nvidia to actually put decent reference cooling on their cards for once.

So yeah, it doesn't seem like an ocd 980 ti will beat a fury x that's also ocd.

I'm not super impressed by 5% performance gain on Fury X OC when a typical 980 Ti OC is like 20% which will both wipe out the Fury X's advantage at stock and also match it at OC. Of course the Fury X will be able to hold it's OC clocks NBD, unlike a stock reference cooler 980 Ti which will slam straight into it's temperature limit and throttle back under constant load. Nvidia needs to do something about their damn reference coolers.
 

Amazing.

From the article:

PC World said:
Speaking of cranking it to 11&#8212;er, 8&#8212;AMD&#8217;s PR keeps stressing that the Fury X will be a kick-ass overclocker. The card&#8217;s design speaks to that, featuring a dual BIOS switch, 6-Phase power design with up to 400 amps of power delivery, and AMD&#8217;s standard SVI2 interface to the voltage regulator, which sports full telemetry readback and lets you tinker with power settings via AMD&#8217;s PowerTune. (If you didn&#8217;t understand any of that, don&#8217;t sweat it&#8212;they&#8217;re hardcore overclocking features.) And while the Fury X typically draws just 275W of power while gaming, the dual 8-pin connectors support up to 375W. Read: OVERCLOCK ME.

Wow, it's gonna be a beast of an overclockable card.
 
Yeah I'm thinking Nvidia always going cheap with their reference coolers is actually going to be a problem really soon now. I hope this actually motivates Nvidia to actually put decent reference cooling on their cards for once.



I'm not super impressed by 5% performance gain on Fury X OC when a typical 980 Ti OC is like 20% which will both wipe out the Fury X's advantage at stock and also match it at OC. Of course the Fury X will be able to hold it's OC clocks NBD, unlike a stock reference cooler 980 Ti which will slam straight into it's temperature limit and throttle back under constant load. Nvidia needs to do something about their damn reference coolers.

Well, it's only a 100mhz OC, I imagine it will do more than that. AMD releasing ocd benches at all is actually a good sign, as it's not something they or Nvidia typically do.
 
Yeah I'm thinking Nvidia always going cheap with their reference coolers is actually going to be a problem really soon now. I hope this actually motivates Nvidia to actually put decent reference cooling on their cards for once.

Eh, I don't feel like either side has ever put too much money into their reference coolers (outside of the 295x2 and the Fury X)... though that said, I've always loved the way nVidia's reference coolers looked haha.
 
So someone help me understand because I feel like the benches show that it isn't that great.

Are the Fury X benches good?

So far with what they have shown, I would say yes. Not to mention that hopefully as more drivers come out of the pipeline performance will improve like it did on the 290x.

I think the Fury X is gonna give the Titan X a run for its money. But then there's the whole NVIDIA gameworks, and NVIDIA sponsored games that will run poorly on AMD hardware (*cough ACU*).
 
So far with what they have shown, I would say yes. Not to mention that hopefully as more drivers come out of the pipeline, performance will improve like it did on the 290x.

I think the Fury X is gonna give the Titan X a run for its money. But then there's the whole NVIDIA gameworks, and NVIDIA sponsored games that will run poorly on AMD hardware (*cough ACU*).
Doesn't Shadow of Mordor favor AMD hardware?
 
So I think I might do a fury x build, and keep the 980 ti build as well. One for the monitor and one for the 4k TV. Sometimes I feel like I'm crazy with all this technology.
 
So far with what they have shown, I would say yes. Not to mention that hopefully as more drivers come out of the pipeline performance will improve like it did on the 290x.

I think the Fury X is gonna give the Titan X a run for its money. But then there's the whole NVIDIA gameworks, and NVIDIA sponsored games that will run poorly on AMD hardware (*cough ACU*).
The VRAM size and driver FUD are going to echo hard in the interwebs.

This card fits my own use case almost perfectly though, all I'm waiting for are stock availability and, as Durante suggested, frametime results.
 
Doesn't Shadow of Mordor favor AMD hardware?

What's impressive about it is shadow of mordor will use 6gb of memory using ultra textures, so if the fury x is able to beat the 980 ti using only 4gb, some kind of magic is going on. Could be they aren't using ultra textures though too.
 
Yeah I'm thinking Nvidia always going cheap with their reference coolers is actually going to be a problem really soon now. I hope this actually motivates Nvidia to actually put decent reference cooling on their cards for once.

I wouldn't necessarily agree. The 780 series+ had pretty decent coolers. They had built in vapour chambers so they were very good even as a blower format. In the 9xx series they dropped the vapour chamber, because the lower TDP didn't require it.

Regardless, even the the lesser cooler on the 9xx is better than AMD's old reference (do they have new reference?).
 
Am I the only one excited about the possible low to no noise coming out of the Fury X card? With the awesome low temps?
Coming from a GTX-680 will be a dream. Just need to wait for real reviews and see if I go Fury X or watercooled 980ti

Pretty exciting year with so many new toys :)

You are not the only one who is excited. The noise issue is one of the reasons I am eagerly awaiting some independent benchmarks on the Fury X. If it's as good as it appears with that nearly silent operation then that is a HUGE plus for me.
 
I'm not super impressed by 5% performance gain on Fury X OC when a typical 980 Ti OC is like 20% which will both wipe out the Fury X's advantage at stock and also match it at OC. Of course the Fury X will be able to hold it's OC clocks NBD, unlike a stock reference cooler 980 Ti which will slam straight into it's temperature limit and throttle back under constant load. Nvidia needs to do something about their damn reference coolers.

My 980 Ti G1 stock is already 20% over reference boosting 1291 on air from the get go and it was only $690. I haven't even bothered taking it further but I will eventually. Techpowerup pulled 1512 out of theirs so I have hope. The Fury X is going to have to overclock like a motherfucker to keep up. It would have been nice to buy a water cooled card stock but I've already invested a considerable sum in Gsync displays and Freesync is still sorely lacking.
 
What's impressive about it is shadow of mordor will use 6gb of memory using ultra textures, so if the fury x is able to beat the 980 ti using only 4gb, some kind of magic is going on. Could be they aren't using ultra textures though too.
The Mordor ultra texture setting turned out to be mostly hype iinm. The 3GB 780Ti still outperformed 6GB Titan in min FPS.
 
The VRAM size and driver FUD are going to echo hard in the interwebs.

This card fits my own use case almost perfectly though, all I'm waiting for are stock availability and, as Durante suggested, frametime results.

Ah yes of course. Aren't most AMD flagship cards known to deliver lots of frames but doing so inconsistently?

I can't/couldn't really tell with my 290x, but in some titles, frame-pacing felt like a huge issue coming from an NVIDIA card. I dunno.
 
Yeah I'm thinking Nvidia always going cheap with their reference coolers is actually going to be a problem really soon now. I hope this actually motivates Nvidia to actually put decent reference cooling on their cards for once.



I'm not super impressed by 5% performance gain on Fury X OC when a typical 980 Ti OC is like 20% which will both wipe out the Fury X's advantage at stock and also match it at OC. Of course the Fury X will be able to hold it's OC clocks NBD, unlike a stock reference cooler 980 Ti which will slam straight into it's temperature limit and throttle back under constant load. Nvidia needs to do something about their damn reference coolers.

My 980 Ti G1 stock is already 20% over reference boosting 1291 on air from the get go and it was only $690. I haven't even bothered taking it further but I will eventually. Techpowerup pulled 1512 out of theirs so I have hope. The Fury X is going to have to overclock like a motherfucker to keep up. It would have been nice to buy a water cooled card stock but I've already invested a considerable sum in Gsync displays and Freesync is still sorely lacking.

There's a difference between OC % in terms of clock speed and actual performance increase. It isn't a 1:1 scale. AMD is showing a 5% framerate increase with 100mhz overclock. Not a 5% clock speed increase.
 
Ah yes of course. Aren't most AMD flagship cards known to deliver lots of frames but doing so inconsistently?

I can't/couldn't really tell with my 290x, but in some titles, frame-pacing felt like a huge issue coming from an NVIDIA card. I dunno.
Ya AFAIK the big AMD frametime hoo-haa last time was for its then-horribad CF performance, after NVIDIA started a conversation about FCAT.
 
I'm not super impressed by 5% performance gain on Fury X OC when a typical 980 Ti OC is like 20% which will both wipe out the Fury X's advantage at stock and also match it at OC. Of course the Fury X will be able to hold it's OC clocks NBD, unlike a stock reference cooler 980 Ti which will slam straight into it's temperature limit and throttle back under constant load. Nvidia needs to do something about their damn reference coolers.

Yeah that OC isn't too impressive. NV could just refresh the entire Maxwell lineup with higher clocks and bring out a fully enabled 980 Ti for ~5% more performance (on top of 15% or so clock bump) to take back the crown.
 
Yeah that OC isn't too impressive. NV could just refresh the entire Maxwell lineup with higher clocks and bring out a fully enabled 980 Ti for ~5% more performance to take back the crown.

Guys, in no way shape or form is AMD saying you will only be able to get 100mhz OC out of it. They were just providing an example. The card has a 375 watt power limit. I would imagine you can easily get at least another 100-200mhz out of it, if not more. We'll find out soon enough.
 
My 980 Ti G1 stock is already 20% over reference boosting 1291 on air from the get go and it was only $690. I haven't even bothered taking it further but I will eventually. Techpowerup pulled 1512 out of theirs so I have hope. The Fury X is going to have to overclock like a motherfucker to keep up. It would have been nice to buy a water cooled card stock but I've already invested a considerable sum in Gsync displays and Freesync is still sorely lacking.

I was in the market for a new monitor and the 980 ti will pretty much decide my purchase based on sync technologies. What don't you like about freesync when compared to gsync?
 
I was in the market for a new monitor and the 980 ti will pretty much decide my purchase based on sync technologies. What don't you like about freesync when compared to gsync?

The Benq XP2730Z which was basically the only competition for the ROG Swift for months had some bad ghosting trouble. The Acer XG270HU is a lot better apparently.

Freesync also has average handling of frame rates outside the specified refresh rates whereas Gsync is quite reliable. For instance with Freesync going outside the monitor's specified refresh range will result in either vsync being turned back on (integer factor of 144Hz frame rate) or bringing tearing back. Gsync on the other hand just refreshes frames at double, triple or quadruple refresh rates as it dips below 35Hz. For the XP2730Z this threshold was 40Hz which is a little high for my tastes. I'm just really sensitive to judder.
 
Who are going to buy a Fury X before a 980ti or a 390x before a 980? It's not enough that they are on par in performance because nVidia cards overclock better.

AMD lost this gen too. Too bad. :/

ok, at least the AMD cards are cheaper.
 
Who are going to buy a Fury X before a 980ti or a 390x before a 980? It's not enough that they are on par in performance because nVidia cards overclock better.

AMD lost this gen too. Too bad. :/

ok, at least the AMD cards are cheaper.

lul wut. There are zero legit in depth benchmarks out shill.
 
The Benq XP2730Z which was basically the only competition for the ROG Swift for months had some bad ghosting trouble. The Acer XG270HU is a lot better apparently.

Freesync also has average handling of frame rates outside the specified refresh rates whereas Gsync is quite reliable. For instance with Freesync going outside the monitor's specified refresh range will result in either vsync being turned back on (integer factor of 144Hz frame rate) or bringing tearing back. Gsync on the other hand just refreshes frames at double, triple or quadruple refresh rates as it dips below 35Hz. For the XP2730Z this threshold was 40Hz which is a little high for my tastes. I'm just really sensitive to judder.

The ghosting on the BenQ XL2730Z was an overdrive issue which was fixed with a new firmware. New shipments of the monitor won't have that problem anymore.
 
The ghosting on the BenQ XL2730Z was an overdrive issue which was fixed with a new firmware. New shipments of the monitor won't have that problem anymore.

Bully! The more people who can enjoy high quality Adaptive Sync the better! Especially since it's $200 cheaper than the Gsync equivalent.
 
Listened to a little bit of the TechReport stream. The older guy (sorry, don't know his name but he was at the Fury launch and talked to the product manager from AMD) said the HDMI 2.0 thing was not a big deal to him or AMD simply because most hardcore gamers preferred to game on the 144hz monitors at 1440p and all of those use display port.

That's a reasonable answer but it still stinks if you are one of those people who have a 4k TV that only has HDMI.
 
Listened to a little bit of the TechReport stream. The older guy (sorry, don't know his name but he was at the Fury launch and talked to the product manager from AMD) said the HDMI 2.0 thing was not a big deal to him or AMD simply because most hardcore gamers preferred to game on the 144hz monitors at 1440p and all of those use display port.

That's a reasonable answer but it still stinks if you are one of those people who have a 4k TV that only has HDMI.
The older guy is Scott Wasson ;)

I'm guessing you've seen it, but I just caught up on the stream and regarding 4GB they don't have a concrete answer on how much they are doing with driver optimisations. He did though say that software tools for GPU memory usage aren't very helpful so they'll have to do performance tests instead by ramping up memory intensive settings like AA.
 
Listened to a little bit of the TechReport stream. The older guy (sorry, don't know his name but he was at the Fury launch and talked to the product manager from AMD) said the HDMI 2.0 thing was not a big deal to him or AMD simply because most hardcore gamers preferred to game on the 144hz monitors at 1440p and all of those use display port.

That's a reasonable answer but it still stinks if you are one of those people who have a 4k TV that only has HDMI.

Eh, he's right IMO. It's meaningless unless you have a 4k TV... then it's a big issue.
 
Top Bottom