****Official OcUK Fury X Review Thread****

[CAT-THE-FIFTH]

See you're a 100% kickass fan boy
I look at the whole package and what is happening to other users.

Waiting for gibbo's review. I will most likely go back to NV with a TI G1.

Seriously,dude what res are you running ED at??

Everyone I know who runs it,is fine with sub £200 hardware. Its not that taxing for a game with such scope and is well optimised.

 

[CAT-THE-FIFTH]

The issue is when using CF it seems to screw the game up.

Single card it's fine but crossfire introduces flickering and stutter.

So run single card?? Most of my mates are running the game at 1080p with slower cards than a R9 290X/GTX970 and its perfectly fine.

I know a few people with a GTX970 who run the game fine on a single card.

 

[CAT-THE-FIFTH]

Why should they have to run single card though and not use a few hundred quids worth of hardware they've paid for?

If crossfire worked it means they could crank the settings up higher and still have more fps.

It's only one game so it's not a massive issue but AMD should still have addressed it by now.

The problem is that even at decent settings you don't NEED massive graphics horsepower with the game. Its incredibly well optimised - I was half expecting my mates to ditch their cards ,but apparently not. Maybe the chap is running a 12K setup or something? Shrugs.

Plus don't disagree with the XFire profiles issues BTW.

 

[CAT-THE-FIFTH]

You *could* argue that the rig HardOCP used is pretty much an average gamer one (and the one more likely to be owned!)

I do get the impression from your posts you do seem to have a lot of purchase justification going on ATM.

 

[CAT-THE-FIFTH]

LOL. Just lol.

All the hype and anticipation... and it's another crap card from AMD. Disappointing.

AMD should focus on the mid-range instead. That's where they've always done best.

Imagine if they had rejigged R9 290X with the improvements from the R9 285(better tessellation and the memory compression thingy) with HBM??

They might have not been competing with the GTX980TI and Titan X,but they would have had a better GTX970/GTX980 competitor and probably would have dropped power consumption too a decent amount and had maybe even a smaller chip than Hawaii.

That way they could have found themselves more likely to be in laptops,and could have done better on price/performance.

They infact did this with the HD3870 and HD4870 releases and grew marketshare AFAIK.

OFC,it could be since HBM availability is poor so maybe not so easy,but they are loosing in the midrange especially for laptops.

 

[CAT-THE-FIFTH]

Guess I'm looking at it purely from my perspective.

MATX case means I need to use ref cards to exhaust heat.

The AIO cooler does the same, and quieter, so for my specific needs, that's a win.

I would have to agree here,as I only use SFF rigs,and its the graphics card NOT the CPU which produces the most heat. Its why the normal AMD reference blower cooler for the R9 290 series was so meh,especially for more compact rigs.

AMD screw their customers. Look at them hobbling R9 290x performance vs R9 390x by limiting drivers to 300 series only.

Deliberate downgrade to the 290x to encourage upgrades.

Tell that to Kepler customers like me,who are not seeing much driver optimisations for our cards.

Its just wrong that in W3 with Gameworks off,an R9 290/290X was getting the better of GTX780 cards in a Nvidia sponsored game,and that is with AMD not evening releasing a flipping driver for the game.

 

[CAT-THE-FIFTH]

Remind me to sell my Maxwell cards once Pascal comes along

(at least mainstream Pascal!! Big Daddy Titan will come first )

More likely mainstream Pascal will come first.

 

[CAT-THE-FIFTH]

Hmmmmmm, I'm thinking Titan in Jan/Feb then the rest to follow tbh. They'll push HBM2.0 out on something asap to not look to be left behind

Considering AMD is a partner with Hynix on HBM,they will probably be out the boat first anyway IMHO with HBM2 and its still more likely midrange first for retail for Nvidia. They did that for the GTX680 against the Titan and the GTX980 against the Titan X.

Any initial large die Pascal cards will be going for HPC - the GM110 based cards were made available for commercial customers nearly six months before retail customers.

If 28NM is anything to go by costs,its more likely to more of the same with 16NM and 14NM,so the higher profit margin customers will be served first,especially as yields will be another concern and companies like Apple for example will get priority over small players like AMD and Nvidia.

Plus since the GM210 has been out only since March,I don't think Nvidia will be replacing it within a year for retail.

 

[CAT-THE-FIFTH]

This is assuming AMD have the funds/budget to create another GPU with HBM2 in the reasonable future.

Remember how long AMD's top card was the 290X? Remember how many new GPU's NVIDIA managed to launch in that timeframe?

I wouldn't be surprised if FuryX/Fury/Nano/Fury 2X are all we have for another 2 years.

Or maybe they are pushing all their R and D to get as early as possible onto the next process node??

There is a lot of noise that they will be using GF instead of TSMC,meaning they might not be as supply constrained since TSMC has a lot of big players like Apple and Qualcomm using their fabs.

We need to be very careful about too much doom and gloom.

AMD had the 2900XT and HD3870 which was followed by both the HD4870 and HD5870.

In many ways they were far less competitive with the 2900XT and HD3870 than they are now.

 

[CAT-THE-FIFTH]

Got anything to back that up?

Nvidia have made huge wins in recent times. They are the sole supplier for IBM supercomputer designs and the US government has recently announced they will be building the three fastest yet announced supercomputers in existence around NV hardware. That's some big money and mindshare right there.

So AMD got Apple, that's peanuts as we know Apple screw down their supplier margins, plus Mac Pro sales are hardly setting the world alight, it has not helped AMD's bottom line much has it?

Edit: You just need to look at the top 500 to see what people who's opinion counts really think about AMD's compute designs.

http://www.top500.org/lists/2014/11/

I believe AMD went from around 12% to 21% of the entire market in the last few years - this includes workstations too and is where AMD has got most of the share.

However,the main concern for Nvidia will be Intel,who are putting more and more resources into Xeon Phi - not only ones which are add-in cards but potentially socketed ones.

Look at the top supercomputer on the list - it use Xeon Phi and so does the 7th ranked.

IBM,has had its own issues too - it basically divested itself of its fabs and started to license POWER uarch to others. The collaboration with Nvidia is more the fact they have nothing by themselves to take on Intel with compute cards,so again the common enemy is Intel not AMD.

 

[CAT-THE-FIFTH]

Aye, I completely agree that Intel and not AMD are the threat to Nvidia in the professional and HPC compute space. But as I said, NV have made some huge recent design win announcements. http://www.anandtech.com/show/8727/nvidia-ibm-supercomputers

Where are the AMD HPC wins coming from?

I would say in the professional space they are stealing sales - its why they have nearly doubled marketshare(it might be more TBH if you consider they are not massively in the HPC space).

However,the current most efficient supercomputer does have an AMD FirePro S9150:

http://www.green500.org/lists/green201411

The S9150 is Hawaii based. Hawaii is actually quite competitive in performance/watt for HPC tasks,but Nvidia simply has more experience in the area now and thats because prior to GCN,AMD and ATI cards were not that great for compute in the first place anyway.

It will be interesting to see when the new lists are out whether AMD has gained anymore traction in HPC,as until next year its still Kepler cards for Nvidia.

 

[CAT-THE-FIFTH]

Yet, nobody still wants to use their offerings in any recent big builds.

Can you link those market share numbers, I'd be interested in having a read.

I was going by marketshare last year,so OFC have no data from recently:

http://www.forbes.com/sites/greatsp...t-share-can-rise-to-30-after-the-macpro-deal/

http://marketrealist.com/2014/10/amds-professional-graphics-attain-a-milestone-with-apple-deal/

First one said 20% in May 2014 and it appears it says 25% in October 2014.
Another report said 21% at the beginning of 2015.

A few years before that they were closer to 10% AFAIK.

 

[CAT-THE-FIFTH]

@CAT-THE-FIFTH

An interesting Fury X review from Forbes..

http://www.forbes.com/sites/jasonevangelho/2015/06/24/amd-radeon-r9-fury-x-review-amd-at-their-best/

Added to list.

Why bother to make a new card when the 390X (or 290X OC'd with new drivers) is competitive with the 980 at 1920x1080, beats it handsomely at 2560x1440 and absolutely murders it at 4K?

Sorry I missed your comment.

Its more the case,a rejigged Hawaii with HBM and the Tonga tweaks,would make it much more power efficient and due to the smaller size of the HBM controller would have made the die smaller too.

This would make the chip easier to use in laptops and buy AMD some much needed marketshare in that area.

Even if Nvidia had the super high end market,AMD could get a lot of sales in the rest of the market which would help them.

However,after I made my comment I realised HBM might have more limited supply than GDDR5 which might not help.

 

[CAT-THE-FIFTH]

1080p performance is poor to say the least.

LOL,so you posted three videos of W3 a Nvidia Gameworks game??
Compared to the R9 390X it has a performance lead.

You do realise in the Crysis 3 video it is neck and neck with the Nvidia cards,right??

Edit!!

With Hairworks off its not to far off a GTX980TI too.

 

[CAT-THE-FIFTH]

lol there are more games tested, try watching it.

And yeah 1080p performance sucks, if you can't see that you must be blind.

Wow,such EMO DORK RAGE!

But seriously,why is the 4K and 2560 performance within a whisker of the GTX980TI with 6GB of VRAM in two Nvidia sponsered games??

What a horrible performance collapse.

Is Nvidia working on better drivers to improve performance in those videos??

 

[CAT-THE-FIFTH]

It's not me raging and it's not me that's posting BS. Deal with it

You are RAGING - you posted 4 videos and when you realised three did not show what you wanted to know you quickly edited.

So you are admitting you posted "BS" as you described it?? 75% going by your own description.

The fact that you insulting people and swearing means you are annoyed and its psychologically affecting you.

Your the one getting worked up not me.

Do some more RAGING its funny!!

So sweet.

 

[CAT-THE-FIFTH]

I edited nothing and



I don't insult people so who's the one worked up?

Now I suggest you drop the BS and keep OT.

MORE RAGING and swearing.

Such techiness. You really get irritable fast don't you? Bless.

Its really psychologically affecting you.

Want to have the last word then??

Get it in,then you can calm down.

Plus,I am the one updating the OP too.

BAHAHAHHAHAHAHHAHAHHAHAHHAHAHHA!

 

[CAT-THE-FIFTH]

If drivers don't get better then I'm afraid it won't happen. Locking the voltage was the death blow imho. Something wrong if we haven't had Sapphire/MSI release voltage unlocking versions of their software on release.

Fury should have been released at around $50 less than the 980ti. I guess the Fury Pro may be the one to redeem AMD's badly managed launch. If it comes in at £400-£430 then It will sit right between the 980 and 980ti at an enticing price.

Agreed with all of these points.

Such a shame the Fury X didn't turn out to be the monster it was hyped to be.. They really needed to beat the 980Ti consistently across all benchmarks to make a mark, even then they really needed to outperform the Titan X to show a significant gain over Nvidia to put them into a power position.

What is equally disappointing is what this will mean going forward, as i can't see AMD lasting very long now and an eventual sell off is on the cards with recent split rumours. If AMD were to exit the mid to high end GPU market in the next 1-2 years, it wouldn't shock me in the slightest. If they do, we'll be left with Nvidia who can pretty much price their products however they like with no competition.. Bad for all.

Always confuses me why some people celebrate the misfortune of AMD, because it just means a future of more expensive cards from Nvidia.

I'm going to stick with my 290X, let Oculus Rift drop in Q1 2016 and then upgrade to whatever Pascal product lands in Q2/Q3 2016 from Nvidia.. Unless AMD can pull some magic off? It's doubtful, this really felt like their last stab.

OTH,it could be that there realise they were in a fix and have decided to get as quickly to the next node shrink as possible and I suspect they will move to GF instead of TSMC this time too.

We will see.

But TBF I would have no problem if AMD just concentrated on the sub £300 market in a decent way. The HD4000 and HD5000 series might have not been the top dogs once Nvidia launched there stuff but it still worked for the market,especially for mobile.

Nvidia has done well since they concentrated more on the midrange recently - the GK104,GM107 and even to a degree the GM204 were small GPUs by Nvidia standards and they all got a lot of laptop design wins.

 

[CAT-THE-FIFTH]

Added some more reviews to the OP.

 

[CAT-THE-FIFTH]

Nice.

Interesting RAM usage results on this one;

http://www.extremetech.com/gaming/2...erformance-power-consumption-and-4k-scaling/2

Still rather surprised at 4K it does not collapse inwards in performance.

It makes me wonder what is limiting the Nvidia cards - its not the core speed or VRAM amount for sure.

I doubt its bandwidth as the Maxwell cards have colour compression too,and if the AMD chap meant what he said the Fury is doing more texture swapping across the PCI-E bus to the main system RAM anyway.