ASUS STRIX Radeon R9 Fury DC3 CrossFire at 4K Review @ [H]

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,532
ASUS STRIX Radeon R9 Fury DC3 CrossFire at 4K Review - We gear up for multi-GPU gaming with AMD Radeon R9 Fury CrossFire, NVIDIA GeForce GTX 980 SLI, and AMD Radeon R9 390X CrossFire and share our head-to-head results at 4K resolution and find out which solution offers the best gameplay experience. How well does Fiji game when utilized in a CrossFire configuration?
 
Those AMD drivers. :eek:

I'd love to see what the charts would look like if AMD had some NVidia driver people working for them. If there's anything that has ever kept NVidia looking good in the eyes of the public it would definitely be their software side of the house.

Good to see those 290Xs kicking butt, even if they are a ton hungrier than everything else.
 
AMD CF drivers are a real bummer for sure.
Single GPU though, that runs perfectly for me.
 
This is nothing new you can read that AMD does very little for drivers in any [H] crossfire review. The point is that even if it is faster the drivers can (should) be better.

The question is will AMD make a last ditch effort to cut cpu overhead and beef up there crossfire profiles before we get the bulk of games using DX12 ?
 
AMD driver issues are ... one word, LEGENDARY ......

And I think the root cause of these issues is, lack of leadership, real understanding toward their customers and very possibly, the proper resources that it takes, money and software engineers. Why this is an ongoing issue for years is beyond me.

Personally speaking, I finally had enough with Xfire / Tri-Fire 290x's and drivers / performance and moved over to a single 980 ti which has been an utter dream come true.

I'm open to going back to AMD provided they catch up tech wise with HDMI 2.0 and the proper amount of vram for 4K and of course, they finally solve their driver issues. Last but not least, they transition back into a value oriented pricing model like they were before. AMD needs to be cheaper than Nvidia, period. Why? Nvidia does everything right, AMD does not.
 
Nice Review.

I got rid of my crossfire setup for the same issues that you guys are getting.

One video card is much easier to work with. :cool:
 
nVidia's multiGPU is as much fun as AMDs in my experience. The breaking straw for me was when BF4 multiplayer went wonky I sold my second Titan X.

Nice review. I am always interested in how multiGPU is being supported.
 
This was the review I was waiting for. Thanks guys! Crossfire was disappointing for sure. FC4? What the hell, AMD?? Between the Fury and the 390x, I would go with the 390x. HBM showed some promising results, but I agree that 8GB is preferred this generation. Also, $350 is nice to have in the pocket. Between the 390X and the GTX 980, it is a tossup. The 390X seemed to have more potential, but those damn drivers.....
 
Last edited:
Very nice review there.

As always, and as an owner of R9 290X Crossfire and GTX 980 Ti SLi, the biggest problem with AMD is the crap-shoot with drivers.

My 290X Crossfire still has legs, but the driver program at AMD just plain sucks.
I doubt I'll ever own AMD again unless they change their driver support, that or come out with a GPU that's just plain so overwhelming.......well, don't hold your breath.;)
 
Nice review. Reaffirmed my stance on multiGPU issues. One camp doesn't scale and the other camp stutters under DX11. I still hold out hope that DX12's less reliance on drivers will fix this for both camps and ultimately provide a better experience for the sake of the consumers. Going to keep my money in my pocket and not go CrossfireX or SLi. I need less headaches in my life; not more.

Again thanks for the review.
 
No 980 ti SLi in this review. I sometimes really question your decision to include a bunch of cards. If you are spending 1100+change on mgpu why not spend another 200 for a better experience. Showing that is actually a real world decision for many. Oh well. This whole price bucket restriction is really keeping your reviews from going that extra mile in terms of their usefulness.

I really hope next week you put all the charts in one place.
 
No 980 ti SLi in this review. I sometimes really question your decision to include a bunch of cards. If you are spending 1100+change on mgpu why not spend another 200 for a better experience. Showing that is actually a real world decision for many. Oh well. This whole price bucket restriction is really keeping your reviews from going that extra mile in terms of their usefulness.
Well if they include 980 Ti's then they also have to include Fury X's... and then suddenly it's a Fury X CF vs 980 Ti SLi showdown, featuring regular Furys for good measure.
They had the 980s for Nvidia's closest competitor, and the 390X's for a Hawaii comparison. It's fine.
 
i know this doesnt fit into the scope of the review, did you guys by chance try to unlock the cards? just wondering how high or low the success/failure rate is on that.
 
Well if they include 980 Ti's then they also have to include Fury X's... and then suddenly it's a Fury X CF vs 980 Ti SLi showdown, featuring regular Furys for good measure.
They had the 980s for Nvidia's closest competitor, and the 390X's for a Hawaii comparison. It's fine.
I meant next week so edited my post. Yes I want fury x also in there. If they already have a 300 dollar difference among cards then it is ok to have all of them together since the next tier is only 200 more. Anyone making the decision will be better informed that way.
 
I meant next week so edited my post. Yes I want fury x also in there. If they already have a 300 dollar difference among cards then it is ok to have all of them together since the next tier is only 200 more. Anyone making the decision will be better informed that way.
The problem is time. [H] repeats every run with every setup in each review. They do not use any previously collected data, even if the same hardware and same drivers and settings were used. Adding even one card adds hours and hours of setup, tweaking, and actually playing the games to get an idea of whether or not it is playable.
 
I meant next week so edited my post. Yes I want fury x also in there. If they already have a 300 dollar difference among cards then it is ok to have all of them together since the next tier is only 200 more. Anyone making the decision will be better informed that way.

I believe the comparisons are generally based on a dollar to dollar.
All three of these cards are about the same cost.
The Ti is way more so it needs to go up against a FuryX.

The problem is as the authors stated, they had a real hard time getting a second FuryX to test with.
I believe that Hard borrowed PC Percpective's FuryX to use for the Crossfire 4K testing against the 980Ti.:D
 
I've not owned an AMD card since the X800 AIW and could not be happier. Their drivers sucked then and still suck now. From the very first benchmarks in this article, we see wonky behavior. Combine that with their general instability/glitchy nature and you have case for avoiding their products.
 
I've not owned an AMD card since the X800 AIW and could not be happier. Their drivers sucked then and still suck now. From the very first benchmarks in this article, we see wonky behavior. Combine that with their general instability/glitchy nature and you have case for avoiding their products.

but that's certainly for the Crossfire department, in single GPU AMD drivers works perfectly, yes they are slow, extremely slow to release new drivers but they work... and guess what?, they are working for me much better in windows 10 than Nvidia drivers, of course single GPU only.. also remember, that very first benchmark, (project cars) was a game that when launched a GTX 560 was able to outperform a R9 280X by a wide margin or a GTX 960 performed even better than a R9 290X even at 2560x1440. and i'll say again, in my experience with A LOT of Nvidia GPUs in windows 10, AMD have better single GPU drivers.
 
I had dual Asus DirectCU Radeon 6970's back in late 2011/early 2012, and it seems like the type of issues I had then are still present.

The Dual 980ti's I have now are not perfect by any means. I would still argue that if a single GPU is fast enough for you, it is by far a better solution, but they do appear to have fewer issues in SLI than AMD does in Crossfire.

Regarding stuttering, where the article states it is clearly visible in game, but doesn't show up on charts, I wonder if graphing the frame time in addition to the framerate would help reveal this? In a single GPU setup, the frame time should be exactly 1/fps, but in multi-gpu setups this isn't necessarily the case unless you are running with vsync on with low GPU load. In my experience I can usually quantify the stutter in terms of frame time.

The funny thing is that my dual 980ti system really heats up my office after an extended gaming session. Seeing how much more heat the other solutions generate, I can only imagine what that is like!
 
but that's certainly for the Crossfire department, in single GPU AMD drivers works perfectly, yes they are slow, extremely slow to release new drivers but they work... and guess what?, they are working for me much better in windows 10 than Nvidia drivers, of course single GPU only.. also remember, that very first benchmark, (project cars) was a game that when launched a GTX 560 was able to outperform a R9 280X by a wide margin or a GTX 960 performed even better than a R9 290X even at 2560x1440. and i'll say again, in my experience with A LOT of Nvidia GPUs in windows 10, AMD have better single GPU drivers.

Agree.

There are still issues in Windows 10 with some Nvidia setups.

Every time I get a little game time, I struggle with if I want to spend the time restoring my Win 8.1 image, or just play and deal with the issues. Thus far getting some game time in has won.
 
Great review, it's sad to see some of the crossfire bugs going on for AMD. On a positive note the R9-390x seemed to hold it's own and looked pretty good for the money.

One thing that really bugged me about this review, was having to read the full name "ASUS STRIX Radeon R9 Fury DC3 CrossFire" over and over again on the pages for game reviews. Maybe all the reviews are like this, but normally it doesn't stand out so much to me. Possibly shorter named cards? :D AMD Radeon R9 390X CrossFire and GeForce GTX 980 SLI wasn't as bad to read, but having to read ASUS STRIX Radeon R9 Fury DC3 CrossFire in what felt like every other sentence was a bit exhausting.

I don't know if Asus pays you per mention in an article, but if they did, they would owe you guys a lot of money right now :D
Work is slow today so I counted up how many times ASUS STRIX Radeon R9 Fury DC3 CrossFire appeared on the game review pages, (not including the graphs, only text from the review, I did count the title though, so that adds 1 to each page).

14x project cars -14x the witcher 3 - 15x gta v - 19x dying light - 2x far cry 4 (I took one off because it didn't have Crossfire) - 6x battlefield 4 = 70x total.
At least there shouldn't be any confusion about which cards you were referring to in the article
 
Well, they never fixed the flickering in CrossFire while playing Titanfall so I don't hold out much help. I have two 280xs and Crossfire issues are a fact of life. Some games it just doesn't work properly or the performance is worse (E.G. Guildwars 2). I've been playing Skyforce lately and it won't work properly even with Crossfire disabled unless I play in a borderless window, my Geforce 650m-equipped laptop has no issues.

But on the plus side I find AMD's drivers are actually more stable than Nvidia's. I really never get driver related crashes on my desktop but I do occasionally on the laptop and it's not heat-related because it's an Alienware with upgraded cooling.
 
Well, they never fixed the flickering in CrossFire while playing Titanfall so I don't hold out much help. I have two 280xs and Crossfire issues are a fact of life. Some games it just doesn't work properly or the performance is worse (E.G. Guildwars 2). I've been playing Skyforce lately and it won't work properly even with Crossfire disabled unless I play in a borderless window, my Geforce 650m-equipped laptop has no issues.

But on the plus side I find AMD's drivers are actually more stable than Nvidia's. I really never get driver related crashes on my desktop but I do occasionally on the laptop and it's not heat-related because it's an Alienware with upgraded cooling.

you can't take seriously Xfire anything lower than R9 290 (due XDMA) Tahiti works horribly with Xfire.
 
Great review, it's sad to see some of the crossfire bugs going on for AMD. On a positive note the R9-390x seemed to hold it's own and looked pretty good for the money.

One thing that really bugged me about this review, was having to read the full name "ASUS STRIX Radeon R9 Fury DC3 CrossFire" over and over again on the pages for game reviews. Maybe all the reviews are like this, but normally it doesn't stand out so much to me. Possibly shorter named cards? :D AMD Radeon R9 390X CrossFire and GeForce GTX 980 SLI wasn't as bad to read, but having to read ASUS STRIX Radeon R9 Fury DC3 CrossFire in what felt like every other sentence was a bit exhausting.

I don't know if Asus pays you per mention in an article, but if they did, they would owe you guys a lot of money right now :D
Work is slow today so I counted up how many times ASUS STRIX Radeon R9 Fury DC3 CrossFire appeared on the game review pages, (not including the graphs, only text from the review, I did count the title though, so that adds 1 to each page).

14x project cars -14x the witcher 3 - 15x gta v - 19x dying light - 2x far cry 4 (I took one off because it didn't have Crossfire) - 6x battlefield 4 = 70x total.
At least there shouldn't be any confusion about which cards you were referring to in the article

Well that is something you would not expect to hear any day since the launch of those cards were met with: rebrand, old tech & bios hack ...

Don't forget that most of the audience can read past this (I hope :) ) and for [H] to make money on this is cool with me :)
 
They do not use any previously collected data, even if the same hardware and same drivers and settings were used.

Actually that is incorrect. We will use previous data... IF ... drivers have not changed and the game version/game patch has not been changed. Honestly though, the times that happens is few and far between.

No 980 ti SLi in this review. I sometimes really question your decision to include a bunch of cards. If you are spending 1100+change on mgpu why not spend another 200 for a better experience. Showing that is actually a real world decision for many. Oh well. This whole price bucket restriction is really keeping your reviews from going that extra mile in terms of their usefulness.

I really hope next week you put all the charts in one place.

Our usual focus is to compare cards that have price parity, or something close to it. That is a first goal with cards included, sometimes we reach it, sometimes we do not. Brent and I also made a conscious decision to break this into two articles simply because a the amount of data shown.

Feel free to hit that red X up in the corner, it will solve all your issues with HardOCP and its review format.

I believe that Hard borrowed PC Percpective's FuryX to use for the Crossfire 4K testing against the 980Ti.:D

Yes, Ryan loaned us one as we were having a hard to time buying one in terms of finding stock at the time we needed it. I waited for about a week and never could get my hands on one that was not marked up well over MSRP.

Zarathustra[H];1041879960 said:
Regarding stuttering, where the article states it is clearly visible in game, but doesn't show up on charts, I wonder if graphing the frame time in addition to the framerate would help reveal this?

Oh yes, frametime charts would show exactly that. But here are the facts around that. We already know it is stuttering and impacting the gaming experience; I do not need a data point to tell me that. Taking frametime data is a super resource intensive task and we simply see no need for it. I do not know data to tell me what the gaming experience will tell me. Your opinion may of course differ, but ours is not going to change on this point.
 
Great review, folks.
And on a related note I like the addition of Project Cars to the lineup. Racing games are always interesting IMHO, because with the exception of flight sims, there is more going on in an image than FPS etc due to speed and the sheer number of THINGS that get shown in a small amount of time. Though at the same time the devs tend to cheat sometimes on certain things for the very reason that they are flying by and you'd never notice. Makes for a good addition to the suite since it does things differently.

I know you guys have some communications with the Folks over at AMD (and not just their sense-stunted PR dept.)... have they ever given you any feedback in regards to drivers? Didn't they have a PR-push sometime last year about getting better at it?
 
Does the stupid Logo continually flash like on the 980 I have? How about the two white LEDS on the power plug ins? Pepper and Fly Shit but it bugs me, get with it ASUS. The only way to turn of the logo is in GeForce Experience and I don't run that.
 
I know you guys have some communications with the Folks over at AMD (and not just their sense-stunted PR dept.)... have they ever given you any feedback in regards to drivers? Didn't they have a PR-push sometime last year about getting better at it?

http://hardocp.com/article/2014/12/08/introducing_new_amd_catalyst_omega_driver

There is takin' the talk, and there is walkin' the walk....two totally different things. Unless the talk manifests itself in the walk, it is all just hot air. It is obvious to me after this last week that AMD is not making CrossFire performance a priority.
 
I dont care abaut this big cards, but out of curiosity , how many of the games that H test now are Nvida sponsored ?

Are they on diffrent game engines and so ?
 
but that tells the readers that.

No, our extensive use of the English language tells our readers that. Should you require frametime data, you will need to go to another site for that specific information.

Any other posts in this thread as to what the article does not cover will be deleted. I am not going to let a few drag this discussion off-topic.
 
I dont care abaut this big cards, but out of curiosity , how many of the games that H test now are Nvida sponsored ?

Are they on diffrent game engines and so ?

4/6 except for GTA V and Battlefield 4. AMD only has one game which is Battlefield 4. Project Cars is notoriously bad on AMD cards due to its forced PhysX implementation; not sure if it's still happening with latest patch.
 
Last edited:
Project Cars is notoriously bad on AMD cards due to its DX11 overhead

I fixed your quote and this has (for the most part) been fixed by AMD and a high end CPU.

Low end CPU's still tank performance with AMD cards in the game. It is 100% CPU and driver related.
 
That's the article I was thinking of. Looking back I had such high hopes that they were getting serious about drivers. As far as I can tell, nothing has really changed from team red. Solid hardware and mediocre software support.

I remember reading that and also thinking it was a sign of good things to come. Then there was three months of silence from AMD on the driver front. I really wish someone from inside of AMD or a former member of the driver team would bring forth some information to explain just what the hell goes on behind the scenes to cause these issues to either never get fixed or keep popping up. I'll give AMD some credit and say their multi-GPU support has gotten better over the years, but considering it started out as atrocious it really couldn't go anywhere but up.
 
Nice article. Definitely time to hit AMD hard on the Crossfire issues. Seems like this is just inexcusable at this point.
 
I am really surprised that your general response to any valid suggestion for reviews has become, please leave the site. You have a review of fury which is a 569 card and are including a 400 dollar card for comparison yet when suggested to include cards from the 650 bucket which according to your criteria fit better as comparisons, you start lashing out on your long time readers.

Have it your way but at least stay consistent with your own made rules and try to show courtesy to your readers in the same way they are showing to you.

I don't know if you read the introduction of the review, but I stated clearly the next part includes Fury X CF, 980 Ti SLI, TITAN X SLI. Feel free to take the data from that review and compare it with this review since the setup is the same, and you will have your answer.

Putting 6 cards comparison in 1 review would have been extreme confusing and too busy to understand anything, the review would be double the size and take double the time and not make a lot of sense. Separating the articles into two articles better helps focus on the important bits.
 
I am really surprised that your general response to any valid suggestion for reviews has become, please leave the site. You have a review of fury which is a 569 card and are including a 400 dollar card for comparison yet when suggested to include cards from the 650 bucket which according to your criteria fit better as comparisons, you start lashing out on your long time readers.

Have it your way but at least stay consistent with your own made rules and try to show courtesy to your readers in the same way they are showing to you.

If you read the review, you know our exact reasoning behind our thoughts on how we laid this out. I am very sorry that you do not like the way we do things. If I have offended your sensibilities in any way, please accept my humble apology for lack of interest not to structure our content to your exact needs.

To address my "own made rules."

(1) Absolutely NO FLAMING, NAME CALLING OR PERSONAL ATTACKS, NO TROLLING. Mutual respect and civilized conversation is the required norm.

No attack was made. Your point was noted, and it was plainly explained to you that it would not happen, and the reasoning behind exactly why. If you do not like our decisions, reasoning, or both, I suggest you look elsewhere for your GPU content.

Now can we keep the conversation on what the review is about instead of what you think it should have been about exactly? Thank you so very much.
 
Last edited:
Good read. I've been wondering if AMD had improved Crossfire drivers since Fury relaese -- apparently not.

Thing is, I played Witcher 3 for a couple hours tonight on my 295x2 and am not seeing stuttering, at least, definitely not what's described in the article. I wonder if the issues are somewhat relegated to the newer cards.

Looking forward to part II.
 
Back
Top