AMD Radeon R9 Fury X Reviews

[Rys]

@silent_guy said:

It's not that bad. The average power consumption from the reviews in the OP is only 10% above the 980 Ti and performance is roughly the same.

I was looking at the TPU (http://www.techpowerup.com/reviews/AMD/R9_Fury_X/32.html) and hardware.fr numbers. Those have the 980Ti about 20%-33%) ahead.

It's come a long way from the 290X/390X (and yes, its largely due to HBM), and is within a margin now where I doubt it will affect potential purchasers.

I agree: with a water cooler, it's really not a big deal for purchasers. But it shows how AMD has made almost no progress in 3 1/2 years in terms of core power efficiency (other than HBM, of course.) If we add the 50W or so that HBM saves on the full board, the perf/W would drop right back to other GCN cards.

 

[Rys]

@homerdog said:

So a bit slower than the 980Ti for a bit more power. Guess it's still an improvement for them but the price is not right (Why buy a slower, more power hungry card when the 980Ti is the same price and has %50 more memory to boot?). Would make sense at 00.

 

[Rys]

@silent_guy said:

Well after the 750 ti was launched, they should have had a good idea of what was coming up, but at that point it might have been already too late.

They were probably just a couple of months away from tape-out when the 750Ti was released. That must have been an unpleasant surprise, but denial was still an option. Gm204 must have crushed the last sprinkles of hope.

 

[Rys]

@silent_guy said:

Crude Summation:

Site 980 Ti Fury X​

Toms Hardware 233 221
Overclock3D 443 467
TechReport 330 380
HardwareFr 223 267
Guru3D 250 294
PCPer 365 383
TechPowerUp 211 246
PCGH 235 321
HardOCP 372 395
HardwareCanucks 388 403
Forbes 378 411
HardwareBG 365 400
BitTech 422 445

Average 324 356​

You're looking at absolute power and mixing board power and system power. You have to look at board power only and absolute performance if you want to evaluate perf/W. Hardware.fr and TPU give you exactly that.

 

[Rys]

@kalelovil said:

You're looking at absolute power and mixing board power and system power. You have to look at board power only and absolute performance if you want to evaluate perf/W. Hardware.fr and TPU give you exactly that.

Yes, my mistake.

 

[Rys]

@Alexko said:

So a bit slower than the 980Ti for a bit more power. Guess it's still an improvement for them but the price is not right (Why buy a slower, more power hungry card when the 980Ti is the same price and has %50 more memory to boot?). Would make sense at 00.

The quietness of the water-cooling system can matter a lot to some people, but I think that for most, Fiji is 0~100 too expensive.

The cut-down, non-X Fury might be more interesting, though. There's a good chance that a few disabled CUs will barely affect performance, and it should be 00 cheaper. It won't, however, be water-cooled.

 

[Rys]

@homerdog said:

Personally I would rather get a card that doesn't require watercooling to be (almost) competitive, but I guess some folks go in for that stuff.

Also a watercooled 980Ti will overclock like the dickens to vastly outperform the FuryX. Sure it would cost more but you get what you pay for.

 

[Rys]

@pharma said:

From the perspective of an overclocker and gamer:

Sure i really was expecting much more out of the Fury X , i am really disappointed at the performance :/ , the lack of voltage control , locked HBM memory overclock is really hard to overlook. Even my overclokers friends at hwbot and mad enginners have not been able to bypass this , not even my close friend Peter Tan ( Shamino ) has been able to give us a workaround around this overclocking problems , bios seems to be locked too , so no modifications for now until the hard core hard volt mods soldering VRs to this small pcb can be made and / or find a way to unlock the bios. My Overclocker point of view and my opinion , not worth it all the hype and even with volt mods ( if / or when available ) i still have no hopes for Fury X to beat the highly overclockable 980Ti when both are overclocked to the max.

My Gamer opinion , it an ok card for gaming but not as good as the 980Ti , is no brainer really , 6GB > 4GB , 980TI OC > Fury X OC , Performance 980Ti > Fury X. Also the lack of HDMI 2.0 support , no DVI connections and you cannot replace the small AIO liquid cooler are the facts that make the Fury X less appealing for gamers like me with full custom water loops , as it seems there will be no water blocks made for Fury X :/ as it makes absolutely no sense at all ( will have to pray for that ). Coming from a 4x 290X Quadfire set up myself , the lack of drivers and optimizations for more than one AMD card is another big let down , at minimum abysmal crossfire support and is the reality , a fact.

Well as of today my new upgrade path look like this = No more CF or SLI for me , enough is enough. One 980Ti and i will add a full cover water block and overclock it to the moon with a moded bios , Skylake cpu , Socket 1151 Z-170 chipset asus motherboatd , 16GB 3200 DDR4 Ram , and finally new OS windows 10 , so i am not in a big hurry for my big next upgrade and my htpc on my specifications will have to hold me over until then . Well that was my personal opinion as a profesional overclocker and gamer

 

[Rys]

@trinibwoy said:

Geez the [H] review was scathing. Others were more forgiving.

Must've been a sad day in AMD's marketing dept when The 980 Ti dropped for $650.

 

[Rys]

@McHuj said:

Seems disappointing to me overall. 50% more flops and 33% more bandwidth than the 980Ti, yet performance is real close. AMD needs a new and much more efficient architecture.

 

[Rys]

@iroboto said:

Seems disappointing to me overall. 50% more flops and 33% more bandwidth than the 980Ti, yet performance is real close. AMD needs a new and much more efficient architecture.

I'm wondering if drivers could be an issue? The performance might scale well with newer games all designed for directx12/vulkan.

 

[Rys]

@A1xLLcqAgt0qc2RyMz0y said:

The quietness of the water-cooling system can matter a lot to some people

Tom's shows that it is not quiet:

The fan from Nidec almost seems louder at idle than under load at higher RPM. It’s also annoying that the pump and fan generate vibrations that travel along the tubing. How bad these vibrations are depends on the routing and tightness of the tubing. The almost 34 dB(A) produced by the radiator’s fans at idle are a bit too much, especially since the GPU is cool enough. Yet again, the root of the problem can be found in the form of the fan, which produces audible bearing and motor noises.

http://www.tomshardware.com/reviews/amd-radeon-r9-fury-x,4196-8.html

 

[Rys]

@bdmosky said:

I think the regular Fury will be interesting just to see if AMD had to go with this cooler to make it competitive at all.

 

[Rys]

@CarstenS said:

You're looking at absolute power and mixing board power and system power. You have to look at board power only and absolute performance if you want to evaluate perf/W. Hardware.fr and TPU give you exactly that.

For what it's worth:
We (PCGH) are measuring isolated board power as well. For gaming loads, we use two scenes in Anno 2070 (in anglo-american, i think it's marketed as DawnofDiscovery or sth.) as well as Risen 3. Those scenes give a very high load while being very reproducible. It's kind of worst-case-yet-gaming.

 

[Rys]

@McHuj said:

I'm wondering if drivers could be an issue? The performance might scale well with newer games all designed for directx12/vulkan.

That's interesting. If I vaguely recall, AMD cards benefited more with DX12 so maybe they can still see more gains.

Anyone see any BF4 Mantle benchmarks out there?

 

[Rys]

@Babel-17 said:

As availability improves, and AMD gets a handle on tuning the memory and voltages, we might see AMD's AIB partners allowed some latitude. It wouldn't hurt to throw in an upcoming game that will be very demanding as part of a bundle. Anything shader intensive on the horizon?
I think that ultimately AMD will be able to sell as many as they can produce, and without cutting too deeply into their current profit margin. But that assumes they are going to be aggressive with their drivers, tuning, marketing, and helping their partners getting a tweaked version out there. Looks like there might be room for improvement on the overclocking front.

 

[Rys]

@swaaye said:

May be AMD should have gone with more pixel throughput for Fiji this time. There's still leftover BW for that to burn in.

p.s.:

Looks like Fiji's frame-buffer compression is only a bit better than Kepler's:

Now I revise my suggestion -- AMD should have put a more aggressive colour compression instead.

It's strange that it falls behind at lower resolutions. Driver overhead?

It appears to scale very well once resolution is at least 4K. More fillrate would surely help there.

 

[Rys]

@fellix said:

More fragments/pixels, indeed. It just doesn't seem to be able to consume that bandwidth with straight fill-rates, except some mild bottlenecking with FP16 blending.
It is a hell of a texture streamer though, sadly still stuck at half-rate for 16bpc.

 

[Rys]

@kalelovil said:

That's interesting. If I vaguely recall, AMD cards benefited more with DX12 so maybe they can still see more gains.

Anyone see any BF4 Mantle benchmarks out there?

TechReport used DX11 for BF4 on the Fury X since Mantle was showing significantly worse performance.
http://techreport.com/review/28513/amd-radeon-r9-fury-x-graphics-card-reviewed/11

Semiaccurate however reported significant performance gains from Mantle (at 1080p however) in BF4, Civ:BE and Sniper Elite 3.
http://semiaccurate.com/2015/06/24/amds-radeon-goes-premium-with-the-fury-x/

 

[Rys]

@gamervivek said:

Saw the HardOCP review first considering they had a rather good one for 390X and

The saving grace for AMD might be if they are still struggling with drivers for a drastically different gpu setup than Hawaii and get that sorted out by the time they launch the 8GB version(yes that too). Doesn't look like 980Ti's equal.