AMD Fury Series is here. Discussion thread.

frunction

Limp Gawd
Joined
Jan 25, 2013
Messages
214
As requested, thread for discussing the AMD Fury X now that it has arrived and been reviewed.

Personally, my care factor has gone down considerably.
 
Early reviews paint an underwhelming picture for the Fury X, but I feel like things should change with new drivers. I personally think they should have released an air-cooled version first. Releasing one second is kind of backwards logic. With the underwhelming Fury X reviews, my expectation for the air-cooled Fury are rather tepid. The Nano could still be an interesting card, especially for SFF, price pending.
 
Im going to wait a month or more and see some DX12 benchmarks, and hope that a Fury X Ultra OC Edition comes out and at least Beats the Titan X by 5% before i consider buying one.. Would also like to get one for $500 or less if possible

I was super hyped but AMD didnt deliver what i was hoping for.
 
Im going to wait a month or more and see some DX12 benchmarks, and hope that a Fury X Ultra OC Edition comes out and at least Beats the Titan X by 5% before i consider buying one.. Would also like to get one for $500 or less if possible

I was super hyped but AMD didnt deliver what i was hoping for.

What DX12 games are you hoping exist in one month?

Cause I have a feeling there won't be any
 
What DX12 games are you hoping exist in one month?

Cause I have a feeling there won't be any

Who knows, maybe some DX12 benchmarks in Windows 10. Fury X is supposed to thrive in the lover level DX12 stuff.

Im likely to wait till next generation again... i bought a freesync monitor thinking the fury x would be a killer card.. i mean its 48% increased performance over the 290x but i dont really feel that my games lag on the 1440 resolution currently so theres no need to upgrade.
 
Who knows, maybe some DX12 benchmarks in Windows 10. Fury X is supposed to thrive in the lover level DX12 stuff.

Im likely to wait till next generation again... i bought a freesync monitor thinking the fury x would be a killer card.. i mean its 48% increased performance over the 290x but i dont really feel that my games lag on the 1440 resolution currently so theres no need to upgrade.

48% increase performance? Where? While maybe in a few areas yes, but the majority no.

I think the 30-35% increase is more realistic range. While not bad, its not great considering NV offerings.....
 
What DX12 games are you hoping exist in one month?

Cause I have a feeling there won't be any

Bulldozer was gonna be better with Windows 8 and multithreaded games...right?

That's the argument that is trying to be made again it seems.
 
Im going to wait a month or more and see some DX12 benchmarks, and hope that a Fury X Ultra OC Edition comes out and at least Beats the Titan X by 5% before i consider buying one.. Would also like to get one for $500 or less if possible

I was super hyped but AMD didnt deliver what i was hoping for.

windows 10 will be good.
project cars add 30% to 290x there vs win 8
Looking forward the 14 and then 29 of July for win 10 myself.
 
Who knows, maybe some DX12 benchmarks in Windows 10. Fury X is supposed to thrive in the lover level DX12 stuff.

Im likely to wait till next generation again... i bought a freesync monitor thinking the fury x would be a killer card.. i mean its 48% increased performance over the 290x but i dont really feel that my games lag on the 1440 resolution currently so theres no need to upgrade.

It's all because you got this information from AMD.

Guess what, their precanned benchmarks with no AA / AF made FuryX look competitive with the 980 Ti. But with 96 ROPs on the TitanX / 980 Ti vs. 64 ROPs on the FuryX, it doesn't stand a chance at competing on the AA / AF benches.

No DX XXX release will help with a hardware limitation like that.
 
I give up on all of this mess.
I'm riding my 280X to 16nm. Unless there's a delay and we get another arch in-between.

We were supposed to be on 20nm right now. I did say, a year ago, my next GPU would be 20nm or lower ONLY. :cool:
 
I give up on all of this mess.
I'm riding my 280X to 16nm. Unless there's a delay and we get another arch in-between.

We were supposed to be on 20nm right now. I did say, a year ago, my next GPU would be 20nm or lower ONLY. :cool:

Why are you riding out your 280x till 16nm? Werent you up AMD's nuts about Fury being the second coming of Christ? What changed? Its still about 70% faster than your 280x.
 
Why are you riding out your 280x till 16nm? Werent you up AMD's nuts about Fury being the second coming of Christ? What changed? Its still about 70% faster than your 280x.
I was optimistic but I would never put water in my rig.
In its current state or future state it's not worth $550 nor $650.

If it weren't for ramgate I might seriously consider the GTX 970, otherwise I can't justify spending $500+ on another 28nm product.
 
This is the most powerful GPU AMD has ever produced. And people are disappointed. I am confused.

Because for the same price, NV has a competitor that is objectively and subjectively faster, according to all the reviews available. Some sites have it faster than NV's offerings at 4k, but they all agree that no single card runs 4k at any worthwhile settings/FPS so that becomes irrelevant.

Further, AMD made statements to the effect that not only was this going to be a beast of a card but that it was going to overclock like mad. Sure, it has horsepower, but it isnt the biggest or baddest out there, and it apparently SUCKS at overclocking. In short, not only was it not what AMD hyped it up to be, but it cant even top NV's competition for the same price.
 
After letting some reviews sink in, I am still a bit disappointed, but I still think that the Fury X is a solid offering, just priced $50 too high compared to the 980ti. Really, the performance is close, similar power usage and runs cool and quiet. I'll be interested to see how well the architecture holds up over time. The 290/x really seemed to improve over time, so who knows.

But, my wallet spoke and I ordered up a 980ti this time around. Sorry AMD.
 
I guess most of us (open minded ones at least) find it disappointing because we were expecting it to convincingly beat the 980ti and it ended up being slower, not by much though.
It is still a very good card, just not the one that we thought it would be.
 
I'm still going to go AMD for Freesync most likely. The $250 mark up for a G-Sync monitor with lack of inputs bothers me.

My decision has just changed to go Fury X, 295x2, or two 390x. Also, can just wait out for a Fury X2 in fall (but patience is not one of my virtues).
 
I guess most of us (open minded ones at least) find it disappointing because we were expecting it to convincingly beat the 980ti and it ended up being slower, not by much though.
It is still a very good card, just not the one that we thought it would be.

AMD's marketing did not help them in this situation. In that regard, it is like Bulldozer's launch - too much hype and not enough performance out of the gate. Sadly, Bulldozer never matured as we hoped it might, though to be fair a lot of it is the snail's pace of software development/optimization. There is some room for improvement here with driver tweaks and maybe even OCing in the near future, but it is a bit of a letdown at this point.
 
I was waiting for the Fury X and am very disappointed in the reviews. I've never had a problem with AMD drivers, and my last few cards have been AMD. (4850, 4890, 6970, 7950, R9 290X)> Even if I wanted a Fury X my BenQ 144mhz monitor is HDMI and DVI-D only, so I couldn't even play at 144hz. For the price, I see no reason to get this over a 980ti.

I think I'll just stick it out on my 290x for a while longer and see what the next gen brings. I'd rather play 120-144FPS fully maxed at 1080P than 40-50FPS at 4K.
 
The thing I hope people learn from this is to read between the lines a little bit better instead of the whole "OMG WAIT FOR OFFICIAL ANNOUNCEMENT / WAIT FOR OFFICIAL BENCHMARKS" routine. Will save the fanboys a lot of heartache and help preserve a few year on their Doritos clogged hearts.
 
I don't think anyone seriously thought the Fiji chip was going to be a home run, especially after the 20nm fiasco forcing them to stay on 28nm. At most people were hoping it would be slightly faster than the Nvidia counterpart at a much better price. Right now its kinda, ehhh.

I feel sorry for people that actually had high hopes for Fiji, it was just never possible if you looked at all the facts. HBM was never going to be the savior. I think for a lot of people this was needed simple to force Nvidia to lower the 980 Ti to $500, which I think is definitely within the realm of possibilities, so that they could buy that card with no intention of buying AMD.

It is what it is. This Maxwell and Fiji were just stop gaps until 16nm becomes available. Hopefully we get rapid release cycles early Q1 2016. For AMD, Artic Islands could be the actual make or break for them in every sense of the word. Hopefully they get rid of that GCN architecture and do something new. It obviously isn't working out for them nearly as well as what Nvidia has been able to do.
 
These aren't really stop gaps as we're not going to see anything on a node shrink until late 2016 or 2017 and even then, we'll be dealing with unrefined 1st gen stuff versus this node which is probably tweaked to near perfection. Fury X and 980ti will probably have to do for the next couple years before something really good comes along.

As for AMD.. Samsung really needs to buy them out or something to put their GPUs back into competition. They need that supply chain advantage to get an edge.
 
HBM is here to stay. The problem was not enough of it and not enough gpu.
The dual GPU version might have pulled off $650 but this is,not even a $400 card.

As far as HDMI 2.0..it's like female ejaculation. Nice to show off to your friends but that's not needed to have fun with the game.

I am looking forward to next year's version of HBM 2.0 and >14 finfets. I think the die size is limiting both red and green at,this point. Both will be using smaller dies next year along with HBM 2.

And maybe AMD will bring HDMI 2.0 to shut up the early 4k adopters.
I thank those who bought 4K TVs early to pay for the tech but I won't bite till DP is common.
 
I find it comical that they actually don't want you to upgrade to a Fury X if you are running 1440p/1080p.
 
- only 4GB of VRAM
- No HDMI 2.0
- You will have to figure out where to put the radiator
- 649$

Nothing's really going for this card. I wonder what was AMD thinking when they missed out key features. At the current specs it doesn't come close to a 980 Ti let alone beat it on the value proposition.
 
Anybody else notice Fiji XT cores are weaker than Tonga cores?
I used TPUs benchmark, scaled based on the R9 285's performance (1792 cores) and Fury X is anywhere from 10-30% slower than it should be. :rolleyes:

Fiji XT should be at least 128% faster than the 285 based on sp count alone. 4 GB HBM + higher clocks should make it much higher.
I should do Pitcairn cores next I bet it's a closer match.

Hawaii > Tahiti > Tonga > Fiji > ???

Why is GCN getting slower over time? (Via scaling shaders linearly)
 
Wonder if better drivers will make a difference. Maybe they should have delayed this to tune better on day one.
 
- only 4GB of VRAM
- No HDMI 2.0
- You will have to figure out where to put the radiator
- 649$

Nothing's really going for this card. I wonder what was AMD thinking when they missed out key features. At the current specs it doesn't come close to a 980 Ti let alone beat it on the value proposition.

You also forgot:

- Poor overclocking
- Higher power draw
- No DVI (for those who still need it)
 
This thread is so dead compared to all the comments BEFORE the ACTUAL benchmarks came out. It's the first time in a while I can say AMD is losing the price/performance ratio on any series of cards, let alone their "top tier" card.

Just a sad day for AMD. Fury X price cut in 3...2...1...
 
What DX12 games are you hoping exist in one month?

Cause I have a feeling there won't be any

Was going to say exactly this, because I see all over Reddit "just wait til Windows 10, just wait for better drivers, just wait for Air Fury.." Windows 10 isn't the saviour. It'll be years before there's a critical mass of DX12 games, and Nvidia I'll be ready for DX12 too. As well, if the Air Fury performed better they wouldn't have put the water card out first. They put their best foot forward.


Waiting time is over, this was AMD's shot, but went down like someone showing up drunk to their own wedding. You can't blame drivers, AMD had more than enough time. You can't blame locked voltage, they would've had it unlocked for reviews if they didn't think it would fry chips that are already overclocked to the max from factory. There are no excuses.
 
Last edited:
DX 12 hope is as good as Mantle in BF4.

We shouldn't forget what happened to Mantle's performance. After all AMD was all for programming close to meatal.
 
That should be amd fans new slogan, "just wait for ___." It comes and nothing really changes.

Ek's announced waterblock for the fury x looks awesome.
8FG71vr.jpg
 
DX 12 hope is as good as Mantle in BF4.

We shouldn't forget what happened to Mantle's performance. After all AMD was all for programming close to meatal.

You do understand DX12 is vendor neutral? Meaning closer-to-metal in DX12 lifts Nvidia's boat too.

Mantle was AMD only.
 
Almost everything has been said about this card and guess what, I still bought it :eek:

Reason for doing that is PLP-Eyefinity monitor support, which you can't get from the other side even if you sucked JHH's balls dry.

Gonna slap EK block on it and hope that they actually unlock the core voltage at some point.
 
Back
Top