AMD Fury X Reviews

[SimianR]

Yeah - it's definitely not a 980 Ti killer - but it does give it a run for its money in certain games/settings. I think it's a bit of a stretch to say that its real competition is the 980 though. It's performance (at least right now) falls between the 980 and the 980 Ti. You sort of wonder if AMD should have priced is accordingly at $549 (like the 290X) or $599. But sometimes pricing can say a lot - and I don't think AMD wanted to go the "value" route.

I guess it's going to come down as always to what games you're playing and what you yourself value. The 980 Ti definitely seems like the more "balanced" card with 6GB of VRAM and I think it will probably age more gracefully. But having a quiet built in watercooler on the Fury is a nice bonus at its price point.

I think what will truly make or break this product lineup are the R9 nano and the fury (non-X). I'm hoping that taking into consideration the performance of the R9 Fury X, that the Fury (non-X) isn't a cut down chip but full Fiji with a reduced clock speed at most and at a competitive price. That will probably be where AMD competes on price/performance. If the R9 Nano is coming out on August, I'm guessing by then we might already see prices on the R9 390/390X slide down and the Nano could be priced accordingly around $400 (which would probably be the true 980 competition that HardOCP mentioned).

 

[Silverforce11]

The 980 Ti definitely seems like the more "balanced" card with 6GB of VRAM and I think it will probably age more gracefully. But having a quiet built in watercooler on the Fury is a nice bonus at its price point.

Did Titan with 6GB vram age gracefully compared to R290/X with 4GB vram?

If anything, GCN has shown it is the uarch that ages better.

 

[tviceman]

That [H] review is pure shill.

Witcher 3 with HairWorks, check.
FC4 with Enhanced Godrays, check.
Dying Light, check. (ok fine, but when their list of games is tiny, including a few NV biased titles determines the conclusion)

AMD would actually need Fury X to be 50% faster to compete when its crippled so badly.

Then his GTA V result is out of line against almost every other site. As is the BF4 results.

Compare to other sites that use a lot of games, Fury X matches 980Ti at 1440p/4K. It takes a custom 980Ti to beat it (& those custom 980Ti beats Titan X too).

1. There is a Witcher 3 benchmark with hariworks disabled in his review. It's not apples-to-apples, but the 980 TI has higher settings and is still 10% faster. Forget to read the article?
2. How is FC4 fair with Tonga (which does well with Godrays) but suddenly unfair with Fury X?
3. Dying Light has been out long enough for AMD to get performance up. If Hitman was included, which is faster on AMD hardware despite Nvidia having all the time in the world to get performance up-to-snuff, you wouldn't complain one bit.

And just because BF4 and GTA V are skewed doesn't mean he's shilling.

 

[Headfoot]

Talk about Fury X overclocked performance is completely meaningless until we get voltage manipulating utilities.

AMD talked up the overclocking potential of this card. It has massive heat dissipation capability. I'm not going to write off overclocking Fury X yet. It could very well match the 980 Ti overclocks, albeit with extra voltage. It's clearly engineered for overclocking. I would be quite surprised if it doesn't overclock better than a 290x.

 

[VulgarDisplay]

Well at least we now know what Nvidia's response to the Fury X will be. Raising prices on the 980ti. lol

 

[Silverforce11]

1. There is a Witcher 3 benchmark with hariworks disabled in his review. It's not apples-to-apples, but the 980 TI has higher settings and is still 10% faster. Forget to read the article?
2. How is FC4 fair with Tonga (which does well with Godrays) but suddenly unfair with Fury X?
3. Dying Light has been out long enough for AMD to get performance up. If Hitman was included, which is faster on AMD hardware despite Nvidia having all the time in the world to get performance up-to-snuff, you wouldn't complain one bit.

And just because BF4 and GTA V are skewed doesn't mean he's shilling.

Use GameWorks features in NV sponsored games is fair ONLY IF YOU ENJOY less performance on AMD. As an AMD user, I turn that stuff off.

But at least you can comprehend #3. If Hitman was included, hurray... [H] actually has games that favor AMD for once. No really. They have 4-5 games in each review. You can figure out how to skew that result easily enough.

That's why they are useless as a metric on overall performance.

Imagine if [H] added Project Cars to that list.

AMD would literally need a 50% performance advantage in neutral titles to match that kind of sick line-up.

That's why [H] felt entitled to go on an anti-AMD rant. Sure if you play games that clearly favor NV, then Fury X isn't worth it even at $400.

 

[Hitman928]

computerbase did memory testing at 4k, seems AMD wasn't lying about reworking memory management to negate the 4 GB issue on HBM as far as they can test.

 

[bfun_x1]

That [H] review is pure shill.

Witcher 3 with HairWorks, check.
FC4 with Enhanced Godrays, check.
Dying Light, check. (ok fine, but when their list of games is tiny, including a few NV biased titles determines the conclusion)

AMD would actually need Fury X to be 50% faster to compete when its crippled so badly.

Then his GTA V result is out of line against almost every other site. As is the BF4 results.

Compare to other sites that use a lot of games, Fury X matches 980Ti at 1440p/4K. It takes a custom 980Ti to beat it (& those custom 980Ti beats Titan X too).

I'm glad someone noticed that. The Overclock3D review seems a bit more impartial.

 

[Udgnim]

That [H] review is pure shill.

Witcher 3 with HairWorks, check.
FC4 with Enhanced Godrays, check.
Dying Light, check. (ok fine, but when their list of games is tiny, including a few NV biased titles determines the conclusion)

AMD would actually need Fury X to be 50% faster to compete when its crippled so badly.

Then his GTA V result is out of line against almost every other site. As is the BF4 results.

Compare to other sites that use a lot of games, Fury X matches 980Ti at 1440p/4K. It takes a custom 980Ti to beat it (& those custom 980Ti beats Titan X too).

their stance is that it is AMD's job to deal with GameWorks no matter how rigged it's against them

 

[tviceman]

Honestly I would have been OK if Fury X was 5-7% slower max overclocked as at that point it's close enough, but over 20% slower and having 50% less VRAM for almost the same price is just nuts. I don't know what AMD was thinking. During HD4000-5000 era they priced their cards too low and now they went the other way with 390/390X/Fury X pricing where the prices are simply too high. They haven't been able to find this nice middle-ground of price/performance and performance that HD7000 series had.

Maybe AMD can extract 10% more performance with drivers but making up against GM200's 25%+ overclocking headroom is probably not going to happen since 1135mhz Fury X would need to overclock close to 1.4Ghz to keep up without driver increases.

I think another problem AMD is going to be facing is that if finfets are coming until the second half of 2016 as many predict, I think Nvidia will refresh/rebadge their lineup (the 900 series is now nine months old) and the new 1080 TI will be a full GM200 with 5-10% higher clocks. I don't see AMD coming out with a new product lineup until finfets hit.

 

[antihelten]

1. There is a Witcher 3 benchmark with hariworks disabled in his review. It's not apples-to-apples, but the 980 TI has higher settings and is still 10% faster. Forget to read the article?
2. How is FC4 fair with Tonga (which does well with Godrays) but suddenly unfair with Fury X?
3. Dying Light has been out long enough for AMD to get performance up. If Hitman was included, which is faster on AMD hardware despite Nvidia having all the time in the world to get performance up-to-snuff, you wouldn't complain one bit.

And just because BF4 and GTA V are skewed doesn't mean he's shilling.

It is however a bit weird how he concludes that the dying light results must be because of the limited VRAM amount, and then concludes that it must be even worse at 4k:

Sorry AMD, 4GB isn't enough in this game at 1440p, what do you think is going to happen at 4K?

But then when he actually tests 4k, the performance difference between Fury X and 980 Ti, instead of growing even worse as predicted, actually shrinks from 29% down to 11%, but he makes no mention of this whatsoever (as far as I could tell).

 

[SimianR]

Did Titan with 6GB vram age gracefully compared to R290/X with 4GB vram?

If anything, GCN has shown it is the uarch that ages better.

I agree that GCN definitely did age better than Kepler based cards, that's for sure. Who knows how things will play out with Maxwell based cards. Here's hoping that AMD can squeeze more performance out of these cards as time goes on - as I'm sure they will. It's too bad they didn't have that performance now though, because that's where it makes the biggest difference with the first round of reviews.

 

[Hitman928]

 

[AtenRa]

The only thing I saw that I expected is that 4GB is more than enough VRAM even for 4K. Otherwise the performance is meh. The card seems overdone with that watercooling and everything, but still slower than a 980ti at the same price. Disappointing.

I sure hope it will get faster with Catalyst 15.20 on Windows 10, otherwise its not competing in anything, not even in price.

 

[LTC8K6]

Fury appears to be GCN1.2 though?

 

[Silverforce11]

Not only does [H] use mostly NV titles with GameWorks features enabled, their GTA V result has a massive lead for the 980Ti as does BF4. I took a look at toms, Guru, TPU, Computerbase.De and the gap is much much smaller than [H], with 4K basically close-tie.

[H] then blasts AMD for a poor result for Fury X contrary to most other review sites who find it identical to 980Ti at 1440/4K trading blows.

So, along with the vram comments that their own benchmark data proves wrong.. pure shilling site. Removed off my list of tech sites to visit.

ps. I've put many hours into Witcher 3 on my AMD setups, I know full well that it performs a lot better than the results [H] gets.

 

[RussianSensation]

Yeah - it's definitely not a 980 Ti killer - but it does give it a run for its money in certain games/settings. I think it's a bit of a stretch to say that its real competition is the 980 though.

Any site that claims 980 is a Fury X competitor (TR, HardOCP) are just showing their true colours. Brand agnostic reviewers that test a lot of games, not 4-5 games like Project CARS and The Witcher 3 with Hairworks, all show Fury X demolishing a 980 at 1440P and 4K.

If a professional reviewer claims that Fury X is only as good as the 980, they are either getting marketing $ from NV or they are 100% biased in their GPU reviews and simply prefer NV as a gamer or they like doing cherry-picked reviews (for some reason....).

Sites like TR and HardOCP just continue to prove that they rest on their old reputation but today more often than not their reviews are pretty much worthless opinions/drivel against the crop of objective modern European and widely respected likes like TPU and AnandTech. Even Tom's is now better than HardOCP and TR which is mind-boggling. I can pretty much tell if any site is worth its salt by reading their GTX960 review this generation at launch and following launch over time.

It's performance (at least right now) falls between the 980 and the 980 Ti.

That's a bit of a misleading statement. At 4K, 980Ti and Fury X are +/-2 or 3% apart. 980 is so far behind the Fury X, it shouldn't even be in the discussion. About the pricing it at $549-599, I agree.

I think what will truly make or break this product lineup are the R9 nano and the fury (non-X). I'm hoping that taking into consideration the performance of the R9 Fury X, that the Fury (non-X) isn't a cut down chip but full Fiji with a reduced clock speed at most and at a competitive price. That will probably be where AMD competes on price/performance. If the R9 Nano is coming out on August, I'm guessing by then we might already see prices on the R9 390/390X slide down and the Nano could be priced accordingly around $400 (which would probably be the true 980 competition that HardOCP mentioned).

AMD really needs a fully unlocked Fury at $529-549, maybe even $499. That could make it a hit like the 290 was. They should also drop prices on 390/390X. If 390X was $350, it would be an easy recommendation over both the 970 and the 980.

That's why they are useless as a metric on overall performance.

Imagine if [H] added Project Cars to that list.

Exactly, any site that tests just 4-6 games and of these Project CARS and some other GW games are on the list, is a waste of time to gauge overall performance. When other sites test 10-15 games, like TPU, it gives a much more accurate view of the card's overall performance. If someone only plays 4-5 games a year and only AAA games, maybe they find a review with 5 games worthwhile. With sites like TechSpot, Computerbase, TPU, Sweclockers, Hardware.fr, PCGamesHardware busting out such awesome reviews, HardOCP is not even on the map in the list of top tier review sites, especially after their 960 reviews. Let's go back even further and who they praised 660 and 660Ti vs. 7870 and 7950. This was a no contest for any objective PC gamer and now it's evident as much that 660/660Ti get killed by AMD's GCN cards that competed against them. HardOCP though didn't have the brains or the objectivity to recommend buying an HD7950 and overclocking it, getting 3GB of VRAM as a bonus but now they are pimping 6GB vs. 4GB as the big issue. Hilarious double standards that show total lack of professionalism.

It's too bad they didn't have that performance now though, because that's where it makes the biggest difference with the first round of reviews.

They do have the performance at 4K but 1080P and 1440P is where the card begins to struggle. Once we add aftermarket 980Ti into the mix, that's where things get bad for them.

 

[Headfoot]

Fury appears to be GCN1.2 though?

Well no, it has additional features not present in GCN 1.2 so it can't be GCN 1.2. It's different, which ought to be clear by HBM alone...

 

[thilanliyan]

I'm a bit disappointed, but my wallet is ecstatic!!

If they're going to price it the same, they really needed a home run against the 980Ti, which doesn't look to be the case. Hopefully drivers improve the standings.

 

[Headfoot]

AMD either 1) has driver improvements in store, or 2) knows that they need to lower price.

I bet Fury (no x) will be the perf/$ champ here.

 

[exar333]

I think the Fury X needs to be 550 dollars, then it'll be competitive as hell.

Of course I dunno what prices their Nano or Fury pro will be like.

Problem is, it is sort of priced at $550 + $100 CLC = $650.

I agree that it probably should be priced a bit lower than the 980Ti based purely on performance, but the CLC does have value for some. This really makes me think a ref-only option is not the best. A nice 2/3x fan option for a little less would be great too.

Hoping the Pro can OC well and get to a max OC performance close to the Fury X.

 

[monstercameron]

What about compute performance?

 

[utahraptor]

What about compute performance?

From Tom's: http://www.tomshardware.com/reviews/amd-radeon-r9-fury-x,4196.html

On the compute side, Fiji incorporates improved task scheduling and some new data parallel processing instructions to go along with the eight Asynchronous Compute Engines carried over from Hawaii. Given this GPU's 4096 shaders and 1050MHz maximum core frequency, AMD can claim an 8.6 TFLOP single-precision compute rate. However, it limits FP64 to 1/16th of that, yielding a DP ceiling 537.6 MFLOPs (less than Hawaii). After the handicapping of GM200, consider this another nod to the purpose-built nature of high-end gaming GPUs.

 

[Azix]

Don't really think [H] had their heads on straight for that review. Even went off about DVI.

Their conclusions contradict others. eg. Tomshardware looked at getting beyond the 4GB limit and only managed it with settings the cards couldn't handle properly anyway. If you cant get over 4GB without crapping fps whats the point?

I am disappointed but mostly because AMD should know better than to not come out strong. They needed to be clearly better because the market is biased against them. As good as the reference 980ti with less power and quieter just is not enough. I fully expect the card will be better than the 980ti within the year and could be in 2 months if dx12 and windows 10 work in AMDs favor.

Not everybody who buys these things overclock. I would guess most do not. You buy crap like the 980 up for the built in performance, you buy lower because you can reach higher with OC (eg. 970 > 980 with some OC). We'll have to see what the fury overclocks like still.

Fury X = what happens when you slash and burn R&D budgeting and lay off too many engineers.

Also, ouch. Usually HardOCP is an easy grader but... http://www.hardocp.com/article/2015/06/24/amd_radeon_r9_fury_x_video_card_review/11#.VYrEyUbIe4I

So that extra 2.6TFLOPs will rely on ACEs coming into use. If they can get this pushed into games on the PS4 and XBox One then that will help them on the PC side. I would guess nvidia has something similar in maxwell 2 though

On the compute side, Fiji incorporates improved task scheduling and some new data parallel processing instructions to go along with the eight Asynchronous Compute Engines carried over from Hawaii. Given this GPU's 4096 shaders and 1050MHz maximum core frequency, AMD can claim an 8.6 TFLOP single-precision compute rate. However, it limits FP64 to 1/16th of that, yielding a DP ceiling 537.6 MFLOPs (less than Hawaii). After the handicapping of GM200, consider this another nod to the purpose-built nature of high-end gaming GPUs.

 

[monstercameron]

From Tom's: http://www.tomshardware.com/reviews/amd-radeon-r9-fury-x,4196.html

Doesn't say much or show any compute benchmarks.