Right now, rumour is pointing towards AMD finally launching its dual-Fury graphics card, the AMD Radeon Pro, later this week and now that looks even more likely as some benchmark results for the card have leaked on to the web this evening. As we reported last week, at least one retailer has listed the Radeon Pro Duo for release tomorrow and now we have an early idea of what performance to expect against a GTX 980Ti.
Performance numbers for the Radeon Pro Duo were released by Expreview. The graphics card was tested across several modern games including The Division, Rise of the Tomb Raider, The Witcher 3 and Grand Theft Auto V at both 1080p and 4K resolutions.
Being a dual GPU card, performance will often depend on games having decent CrossFire support. That said, the Radeon Pro Duo seems to perform fairly well across these games and by the looks of it, anti-aliasing was left on in each game, which would cause an unnecessary hit to performance at 4K.
If last week's rumours were correct, then we should see the AMD Radeon Pro launch this week, with a price around the £1200-£1300/$1500 mark.
KitGuru Says: AMD's dual-Fury graphics card was unveiled alongside the Fury X way back at E3 last year, which was around ten months ago. It has taken a while but it looks like the card is finally going to hit the market. Do any of you happen to run a dual-GPU rig? What do you think of these performance results for the Radeon Pro Duo?
This is something we need a competition for. I could really do with an upgrade to my GTX 970 about now.
babahahahahahaha what a slow card in a way. i mean its two fury x cards, and yet its only just faster than a 980Ti single card. lol If you place two 980Ti in SLi it will make the duo pro looks like a friggin steam train. hahaha And in less than 4 weeks we will have pascal cards, which will destroy the 980Ti, thus, the math is clear. AMD FAIL big time here. And why would someone buy this overpriced card, when you can buy 2 x fury x single cards and they would be faster than this throttled dual card. just doesnt make any sense for the “crazy price”.
“my room mate Mary Is getting paid on the internet 98$/hr”..,……..!wc174ctwo days ago grey MacLaren P1 I bought after earning 18,512 DoIIars..it was my previous month’s payout..just a little over.17k DoIIars Last month..3-5 hours job a day…with weekly payouts..it’s realy the simplest. job I have ever Do.. I Joined This 7 months. ago. and now making over. hourly 87 DoIIars…Learn. More right Here !wc174:➽:➽:➽➽➽➽ http://GlobalSuperJobsReportsEmploymentsAuthorityGetPayHourly$98…. .❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦::::::!wc174……
But the normal fury X has similar performance to the 980ti….. So I don’t get why you’re saying that at all, this card was also made for professional developers as well hence the different name then what it used to be called because it’s aimed more for developers…….
Not really. 980 Ti outperforms Fury X on most games and once they are overclocked it’s not even a competition. 2x 980 Ti is still better than this.
AMD as always innovating, x86_64 technology, High Bandwidth Memory and now the most powerful video card in the world! Way to go AMD!
So it’s going to be over double the price but offer a lot less than double the performance? Then also rely on decent crossfire support… Disappointing to say the least
at dx12 its a different case though.
At 4K Fury X is in front of 980Ti, at 1440K it’s a tie, only on 1080P or lower 980Ti is ahead. Stock is what it counts to be price comparable. And 980Ti is already much higher in clocks stock-wise…
http://www.techpowerup.com/reviews/Sapphire/R9_390_Nitro/23.html
For 10% Power consumption and 60% more than 980Ti.Fail ? oh you’re Loser!
Scalability across cards is never linear, so twice the cost for less than twice the performance isn’t a disappointment, it’s too be expected.
What’s disappointing is that we have yet to have single and even dual GPU configuration that can push 60 fps at 4 consistently.
Well given it’s price comparison to the 980TI, I see it as very disappointing. Competition wise, it doesn’t really offer any. If someone was looking to upgrade their GPU and was not bothered whether it is AMD or Nvidia, i’m not sure they would come to the conclusion to get this card, given price to performance.
For anything less than 4K gaming, i’d still take a single 980Ti over RPD, much much cheaper and not too big differences in the 1080p scores. Yes 15-40% dependant, but at 1080p all of the scores are above 60fps anyhow.
What’s more interesting is the 4K scores, i’d like to see it compared against 2x 980Ti, as this is like comparing a Titan Z to a single R9 290. Comparing 1 GPU vs 2 GPUs isn’t a fair fight.
All the while though, very nice scores, might move my mATX rig from a single Titan X to a pair of RPDs in the future but 4 way crossfire makes me cringe at the thought of how poor scaling would be and the power needed…
This GPU wasn’t really made for gaming… it is predominantly aimed at compute and professional software use… that’s where all those teraflops are going (and why AMD gpu’s consistently have higher TDP vs Nvidia [which btw is nowhere near AMD in those kinds of tasks]).
We’re comparing Radeon Pro Duo, and GeForce 980 Ti. Both the Radeon series, and GeForce series are gaming orientated. If you’re looking at compute and professional software use, you’d be looking at FirePro and Quadro GPUs.
AMD leaves full double precision support in their gaming gpus, while nVidia keeps double precision mainly in their pro grade cards.
For a cheap compute capable card, Radeon series will do great, but for professional work, you really want to be going for FirePro or Quadro cards as these have more features and drivers that are specifically designed towards compute / professional work.
I have to agree, considering both cards can easily perform over 60fps, 1080p at ultra and I don’t a reason to fork over much more for the Radeon Pro when you can buy a cheaper 980ti and it uses less power
All I’ve seen are rumours for Pascal’s release, I’ve haven’t seen an outright statement from NVidia at all regarding it’s official release, it’s AMD’s new cards that are supposed to be releasing in about 4 weeks I’m sure
Isn’t this a bit of a continuous problem with these benchmarks? The premise that gpus are usually tested and compared based on their gaming capabilities? It gives a woefully inadequate image of a gpus overall capabilities. The radeon pro duo is already going to be used towards analyzing images from space and in VR and it’s been said it wasn’t really targeted at gamers (though it can certainly be used as such with really good results). To solely use gaming as the baseline is ridiculous and short-sighted (which seems to play in nvidias favour).
On top of that, the firepro cards aren’t really different from consumer cards. The only main distinction is in their firmwares and drivers. Amd consumer gpus are a cheaper alternative to far more expensive (and pointless) pro gpus. You could easily make a gpu that does both pro software and gaming in one go, but we’re talking about corporate market here, hence the distinction and price difference.
If you remember, people have been converting their consumer gpus into pro cards through drivers while it was still doable ( both amd and nvidia put a stop to it because they didn’t want people to end up paying less money for the consumer cars that can do the same thing like their pro cards with a simple driver switch – on a hardware level, both are comparable/same/identical with minor alterations to prevent conversion through drivers)
no… if you look they’re really neck and neck theres maybe a 1-5 frame differnce in both of them and only in 1080p the 980ti shines better on most benchmarks as when it comes to 4k the fury x does a better job, yes if you overclock it does make a difference but that’s in any scenario for a card comparing to its counter part in terms of the market point where its listed at. Plus i just looked at scores and the pro duo is getting better performance then 980ti in sli they only differnce is the fire strike ultra scores from what i saw
Fury X as of late offers nearly the same performance as the GTX 980 Ti, and considering that AMD has superior multi GPU scaling, means that the GTX 980 Ti is not going to be able to match two Fury X on CF aka Radeon Pro Duo.
“my room mate Mary Is getting paid on the internet 98$/hr”..,……..!wc263ctwo days ago grey MacLaren P1 I bought after earning 18,512 DoIIars..it was my previous month’s payout..just a little over.17k DoIIars Last month..3-5 hours job a day…with weekly payouts..it’s realy the simplest. job I have ever Do.. I Joined This 7 months. ago. and now making over. hourly 87 DoIIars…Learn. More right Here !wc263:➽:➽:➽➽➽➽ http://GlobalSuperJobsReportsEmploymentsLifeGetPayHourly$98…. .❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦:❖❖:❦❦::::::!wc263……..,…
Can’t really compare a gpu geared for gaming with a gpu geared for compute in a compute test, so they’re compared with what theyre both primarily designed for.
AMD doesn’t necessarily design towards pro users, they leave the options there for compute power at a cheaper price point than FirePro series. There are other factors that differ between the series, such as ECC. We will also likely see a FirePro version of the card, with reduced clocks for better performance / power ratios and all that.
Could almost compare it to comparing an F1 car and a Bugati Veyron,
Both cars are designed to go in a straight line, very fast, while the F1 car is also designed to go round corners, very fast.
In that scenario its like trying to compare the two for their cornering abilities, quite obviously the F1 car will corner much better, but it’s not a fair fight since the Bugati was not necessarily aimed at that.
If nVidia dealt with Prosumer market the same way AMD does, and left full double precision performance enabled in all its GeForce GPUs, then it’d be more equal.
But right now we’re comparing a gaming GPU (980Ti) with a gaming + compute GPU (RPD), its not a fair game at all, the 980Ti likely has somewhere around 150GFlops of double precision integer performance, the RPD may have up to 1/2 of its single precision, so upto 8 TFlops.
If we compared the RPD against say, a Quadro M6000 (same chip as Titan X), then people would be complaining that we’re comparing a consumer grade card with a Pro grade card.
Can’t make everyone happy, just not possible.
But in comparing 2 competing gpu’s, you cannot just pick gaming as 1 thing and recommend the card simply because it may perform better in games while not even mentioning its compute capabilities (especially if its capable of it).
Sure, some gamers might not be interested in compute, but if one is reviewing a gpu and is trying to provide a viable (and unbiased) comparison, you cannot discard a complete portion of the gpu capabilities and say its better or worse overall just on 1 particular feature (gaming) – incidentally, this is exactly what seems to be happening when most websites review gpu’s (they only focus on gaming and nothing else).
The FuryX is not really a professional gpu, though it seems to be targeted at researchers and businesses as a potential cheaper alternative.
So, from a hardware point of view, AMD cards can do both games and pro compute, which I think on balance, makes them a better option.
Sure, if you want to compare just their gaming capabilities, then gaming is what you will focus on… but one shouldn’t use that as a sole means of ascertaining which one is a better gpu overall.
dont wanna wait for new stuff? 2 980ti’s are the competition to this 😛
if you dont compare it to two 980ti’s (which this review should have as that’s its direct competition) yes most powerful gpu 😉
If you believe amd this isn’t for you 😛 But really there are cheaper options on both sides if you dont want to wait for next gen (you should wait)
fury x only outperforms 980ti in most games if it gets help from dx12 in the handful of games actually available under that api.
Yes but how many games support DX12 by the time most do both of these cards (maybe the next even) will be irrelevant for people looking to upgrade.
for more than you could buy 2 980ti hybrids for and have (estimated) 20% less power not overclocked* fixed 😉