Now knowing the performance of Fury X, what would you choose?

[hapkiman]

Thought long and hard (no pun intended), and came to this decision: GTX 980 Ti > Fury X.

Just pulled the trigger on a 980 Ti one day shipping, I'll get it tomorrow. MSI Gaming GTX 980 Ti 6G.

 

[Outback Bronze]

well I will get a month to play with the card, if I don't like it... back to the shop it goes... tbh im more likely to mod the cooling with a better fan at the very least... maybe even switch out the radiator for the one im using on my 290x...

I don't get coil whine as I use vsync when playing games....

if the performance of the card proves to be poor, well then i'll just buy another and another... the joy of not giving a damn and having plenty of cash on the hip.... makes life much more fun...

although if thats the case, this will be the last time I support AMD. you have no idea how much I wanted a 980ti....

btw... if I had gone with a 980 ti it would of arrived this morning...

also if you think the performance is bad then check how it performs in real time in 8 gpu intensive titles. theres nothing in it.

Gees that 390x is very competitive! I wonder how a 290x would fair given its current price. Its got to be one of the best value GPU's out there. Its a pity it chews power like cows on a paddock and heats up like a rocket...

 

[Space Lynx]

Did not vote.


I'd buy a second hand, non-coin farming 290, 290x, or an immensely discounted 7970 GHz. I'd pocket the rest of the money, and invest it toward whichever company can offer better bang for the buck when Pascal goes head-to-head with Arctic Islands. Once everything else was said and done, I'd have a decent card now, and actually be able to afford a card when the die shrink finally comes.

Invest it in stock? Too risky, in which, you hypothetically lose the money, and are stuck with your 290x crossfire with terrible driver xfire support for another 4 years.

 

[lilhasselhoffer]

Invest it in stock? Too risky, in which, you hypothetically lose the money, and are stuck with your 290x crossfire with terrible driver xfire support for another 4 years.

?

I think this might be a linguistic gap. In this instance what is meant is to hold the money now, and whenever the newer cards come out spend it then. There is no plan to try and earn interest or earn dividends to pay for the card in the future.

 

[Space Lynx]

?

I think this might be a linguistic gap. In this instance what is meant is to hold the money now, and whenever the newer cards come out spend it then. There is no plan to try and earn interest or earn dividends to pay for the card in the future.

Yes, this makes more sense. Thanks lol

 

[Warrgarbl]

The 970. In fact, I have bought an Asus GTX970 Strix just three weeks ago. I believe that the GTX970 represents the best price / performance ratio at the moment. I also game on 1.920x1.200, so it is easily sufficient. I'd also rather upgrade the card in 1.5 years or so than keep a much higher priced card for 3 years.

I also really value silence, and this is an area AMD really can't compete in, as much as I do love AMD. I really hope that they will eventually release something awesome again that allows them to put pressure on Nvidia and Intel, especially the former with its monopolist price gouging.

 

[RejZoR]

The memory thing is a huge stinker. Even if it doesn't affect things now, I don't like owning hardware that is known to have something weird going on. That's why I've opted for the GTX 980. One would think NVIDIA has done that intentionally just to sell more GTX 980. Who knows, but still, I prefer to have something that has no "maybe" in it.

 

[GhostRyder]

We aren't talking about the 390X, the Fury X will not have aftermarket coolers, AMD has already said reference designs only.

It definitely isn't whisper quiet, I've heard a Fury X in a case, I just built a system today with one. The rest of the PC is silent at idle, the pump is loud as fuck. Almost as bad as annoying coil whine. Except coil whine is usually covered up by fan noise under load. At idle, the noise is very noticeable. Not to mention when the card is under load you either have headphones on or the sound from the game coming from the speakers to cover up the sound of the computer. At idle that won't be the case a lot of the time. So I'd say quiet at idle is a lot more important than quiet at load. And besides, if we are going by your website, a 3db change is "barely perceptible". So, in other words, no difference.

Again, no it won't be possible on a Fury X card later on. And it definitely is audible when idle.

Where have they said reference designs only because I looked at the slides and around the web and have not seen anything other than the usual at launch there will be only reference designs available (I'm dead serious, not being sarcastic point it out to me). I was pointing out on the 390X since the power consumption is not much different, then its definitely possible for a 0db card/cooler to come out (If not on Fury X, vanilla Fury).

The meters from all sites point the card to being pretty quiet...They do not ignore the sounds the pump makes and report lower levels than what people are hearing. Unless the pump you had was unbelievably loud for some reason they are reported at not being dead quiet like some of the special 0db cards have going but its still quiet especially when load is factored in which is below even most aftermarket coolers on high end cards out there.

I don't see it as anymore audible in a normal case than any graphics card out there. Ill have a chance to experience one though as a friend at a LAN party received 2 of them today for his machine so ill take a listen in person for myself.

 

[purecain]

ok so the radeon fury coil whine is from the pump which looks like a notoriously bad part(its supposed to be fixed on retail samples). im thinking I might just change my order in the morning to a titan x and pay the difference... what say ye?????????????????

but then I read this...
Ms. Premiere: After many months of consideration, I have decided that it is in the best interest of the consumer to know the truth behind the Maxwell architecture. When we were first asked to greatly increase the performance over Kepler at half the power draw, we thought someone upstairs had lost a few marbles. Kepler was already using an advanced form of color compression, transistor gating, variable frequency... we were pushing the limits of 28nm already.



After hundreds of virtual simulations with limited success were conducted, many of the engineers felt that a new approach was necessary. Taking a page from Televisions' book, we began experimenting with interlacing frames, where only half of the frame would be rendered and updated onscreen instead of a whole frame. The approach allowed us to effectively double the frame rates compared to Kepler. In order to properly assign workloads to the GPU, we also were forced to double the number of SM's and halve the number of cores per SM.



This approach saw benefits trickle down into other areas. Using this new approach, we were able to introduce a new form of anti-aliasing which we called Multi-Frame Anti-aliasing, or MFAA for short. In theory, we could obtain double the MSAA strength because only half a frame was being rendered at a time. 2x MSAA suddenly became 4x MFAA. In practice it was more like 2.5x-3.5x due to overhead, but the improvements somehow leaked over to the PR department and they ran with it.



The end result was a new architecture that greatly improved the performance over Kepler, at half the power draw. This lead to many corners being cut in production, including weaker components and Ram partitioning. The original Titan at a thousand dollars was sold at a loss if you include the costs of R&D. The recent Titan X has a 200% markup with all expenses accounted for.

 

[Frag_Maniac]

I believe that the GTX970 represents the best price / performance ratio at the moment. I also game on 1.920x1.200, so it is easily sufficient.

I used to think so, until I started realizing it really isn't if anyone is running considerably higher res or planning to. In fact in extremely VRAM intense games, it only shows as outputting 3.5GB.

That means if you want to run very VRAM intense high graphics games like Shadow of Mordor on the highest settings, you'll be getting VRAM bottle necking and the resulting stutter. And that's just at 1080p.

So at 1080p or anything close to it, most games yeah, but going forward, were I to entertain the idea of a 4K display in the next couple years, or wanted to run all games max just at 1080p even, I'd def go with one of the aftermarket 980 Ti cards, esp the Giga G1. Because the word "value" can start leaving a nasty taste in your mouth once you experince the limitations.

Another way to look at value is what each single card can do in intense 4K gaming, and the Giga G1 almost doubles one 970 in some games at that res, yet you're only using one slot, and outputting the heat of just one card.

 

[R-T-B]

ok so the radeon fury coil whine is from the pump which looks like a notoriously bad part(its supposed to be fixed on retail samples). im thinking I might just change my order in the morning to a titan x and pay the difference... what say ye?????????????????

but then I read this...
Ms. Premiere: After many months of consideration, I have decided that it is in the best interest of the consumer to know the truth behind the Maxwell architecture. When we were first asked to greatly increase the performance over Kepler at half the power draw, we thought someone upstairs had lost a few marbles. Kepler was already using an advanced form of color compression, transistor gating, variable frequency... we were pushing the limits of 28nm already.



After hundreds of virtual simulations with limited success were conducted, many of the engineers felt that a new approach was necessary. Taking a page from Televisions' book, we began experimenting with interlacing frames, where only half of the frame would be rendered and updated onscreen instead of a whole frame. The approach allowed us to effectively double the frame rates compared to Kepler. In order to properly assign workloads to the GPU, we also were forced to double the number of SM's and halve the number of cores per SM.



This approach saw benefits trickle down into other areas. Using this new approach, we were able to introduce a new form of anti-aliasing which we called Multi-Frame Anti-aliasing, or MFAA for short. In theory, we could obtain double the MSAA strength because only half a frame was being rendered at a time. 2x MSAA suddenly became 4x MFAA. In practice it was more like 2.5x-3.5x due to overhead, but the improvements somehow leaked over to the PR department and they ran with it.



The end result was a new architecture that greatly improved the performance over Kepler, at half the power draw. This lead to many corners being cut in production, including weaker components and Ram partitioning. The original Titan at a thousand dollars was sold at a loss if you include the costs of R&D. The recent Titan X has a 200% markup with all expenses accounted for.

You can't seriously believe that... Interlaced frames can actually be visually detected. There is NO way that is going on.

What's been done instead is a cut to double precision to 1/32... The lowest it's ever seen. Fortunately, this is 100% irrelevant to gamers. This is well established.

 

[the54thvoid]

ok so the radeon fury coil whine is from the pump which looks like a notoriously bad part(its supposed to be fixed on retail samples). im thinking I might just change my order in the morning to a titan x and pay the difference... what say ye?????????????????

but then I read this...
Ms. Premiere: After many months of consideration, I have decided that it is in the best interest of the consumer to know the truth behind the Maxwell architecture. When we were first asked to greatly increase the performance over Kepler at half the power draw, we thought someone upstairs had lost a few marbles. Kepler was already using an advanced form of color compression, transistor gating, variable frequency... we were pushing the limits of 28nm already.



After hundreds of virtual simulations with limited success were conducted, many of the engineers felt that a new approach was necessary. Taking a page from Televisions' book, we began experimenting with interlacing frames, where only half of the frame would be rendered and updated onscreen instead of a whole frame. The approach allowed us to effectively double the frame rates compared to Kepler. In order to properly assign workloads to the GPU, we also were forced to double the number of SM's and halve the number of cores per SM.



This approach saw benefits trickle down into other areas. Using this new approach, we were able to introduce a new form of anti-aliasing which we called Multi-Frame Anti-aliasing, or MFAA for short. In theory, we could obtain double the MSAA strength because only half a frame was being rendered at a time. 2x MSAA suddenly became 4x MFAA. In practice it was more like 2.5x-3.5x due to overhead, but the improvements somehow leaked over to the PR department and they ran with it.



The end result was a new architecture that greatly improved the performance over Kepler, at half the power draw. This lead to many corners being cut in production, including weaker components and Ram partitioning. The original Titan at a thousand dollars was sold at a loss if you include the costs of R&D. The recent Titan X has a 200% markup with all expenses accounted for.

FUD. Plain and simple. Pretty sure some of the usual suspects will link dubious articles from shadowy types.
Good thing is, you can read all the reviews, buy an AIB design card and know you'll be far from stock performance and production values.

The Maxwell arch dropped DP compute, that's the saving, as well as other refinements like Fiji has managed over Hawaii.

 

[R-T-B]

FUD. Plain and simple. Pretty sure some of the usual suspects will link dubious articles from shadowy types.
Good thing is, you can read all the reviews, buy an AIB design card and know you'll be far from stock performance and production values.

The Maxwell arch dropped DP compute, that's the saving, as well as other refinements like Fiji has managed over Hawaii.

Actaully, it appears from my brief googling this may be a misinterpreted april fools post on another forum... Either way, NOT true.

 

[purecain]

still what say ye on the titan x instead... ahhg couldn't help myself and went for a Zotac GeForce Titan X 12288MB GDDR5 PCI-Express Graphics Card (ZT-90401-10P).

I managed to get it on special offer at 779.99... cant wait to get the full Maxwell gpu... I wanted this card originally... and I know im not going to be disappointed in any way..


and it will be here on Monday... :0 now im ready for doom on the oculus

 

[GreiverBlade]

still what say ye on the titan x instead...

in a normal scenario i would say not worth it ... but where i live it's funny ... the Fury X range from 800chf~ to 1034chf (with one etailer that does it at 1700ish ... pfahaha so dumb) so it's in line with the price of the 980Ti and Titan X for me, if a 980 was not already enough (and cheaper ... the one i won in the giveaway is 606chf, which is still roughly the MSRP of a 980Ti or a Fury X)

so if i had the money to upgrade around 1000chf i would obviously go Titan X even if its not a card but a trophy since even a 390X and 980 are enough these days, now if the MSRP was respected in Switzerland i would still go Fury X over 980Ti.

the price of the previous Titan line was almost justified due to DP compute performance but the X ... is gimped on that side, so in the end it's just a overpriced gaming card using the name of a previous gen who was a "affordable workstation class card" so now beside the RAM it has nearly nothing to value over a 980Ti for the price asked (heck even a good clocked 980 can be top notch)

oh and coils whine from a pump? well noisy pump in the 1st batch and nope it wasn't corrected in the retail, read the news here http://www.techpowerup.com/213993/amd-revises-pump-block-design-for-radeon-r9-fury-x.html

if it was me i would have already slapped a EK full cover block on the fury i would have, mainly because i already have a GPU loop (hence the "why i am so happy that the one i won was a Poseidon")

 

[husseinHr]

This: EVGA 980 Ti Superclocked ACX 2.0+. Well worth it!

 

[dorsetknob]

Actaully, it appears from my brief googling this may be a misinterpreted april fools post on another forum... Either way, NOT true.

Its Always April fools day in the world of PR

 

[RejZoR]

I think things will drastically change when R9 Fury X2 hits the streets. That one will have 128 ROP's as well as double of everything else. Pack that next to HBM and things will change. Currently I think GPU is gimping the HBM memory...

 

[purecain]

that's why I decided to go with Maxwell. I think NVidia have the better architecture, plus ive never used PhysX.

so im looking forward to seeing what extra particle effects i'll have enabled in the witcher and the new Batman title...

definatly happy with my decision... @hussainhr- too late buddy, I jumped on a titan x reference... i'll watercool it down the road... but thanks for the suggestion much appreciated...

 

[husseinHr]

that's why I decided to go with Maxwell. I think NVidia have the better architecture, plus ive never used PhysX.

so im looking forward to seeing what extra particle effects i'll have enabled in the witcher and the new Batman title...

definatly happy with my decision... @hussainhr- too late buddy, I jumped on a titan x reference... i'll watercool it down the road... but thanks for the suggestion much appreciated...

Ah, right. Well, enjoy the card! It's a monster!