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

Project Cars engine is built around PhysX. AMD has no way to run physx thus their drivers get slowed down to to physx calculations hogging the cpu.

Why would you pick this game for review I have no idea....it's not popular and no one I know on Steam plays it ( 100+ people).

Your information is incorrect.

Please reference - http://hardforum.com/showpost.php?p=1041881127&postcount=76

And - http://hardforum.com/showpost.php?p=1041881116&postcount=75

And - http://hardforum.com/showpost.php?p=1041882104&postcount=111

This game does not use PhysX GPU acceleration.

All 3D features and effects are rendered equally on AMD and NVIDIA GPUs.

Let's stick to the facts please.
 
Felt I should post this here since I did mention the (new) 15.9 beta: http://support.amd.com/en-us/kb-articles/Pages/AMD-Catalyst-15.9-Beta-Memory-Leak.aspx

A little more information (which is all anyone has). The graphics memory will fill with junk. In some games it is not a problem--but the game does show as using all of the available memory. Apparently in some browsers if you resize the image often enough you can produce a crash.

Do I expect this problem to be found and fixed fairly quickly? Yeah, genuinely new bugs can usually be traced to a (small) change made in the new driver.

But again, it may make sense for [H]ardOCP to use beta drivers. You and me? Stay with the stable, endorsed driver (15.7.1) for now.

Beta drivers have the advantage of including new CrossFire profiles and is the only way to get them. If you want the most current game support for CrossFire, Beta drivers are a must.

For Single-GPU beta drivers also include new game support or tweaks or performance optimizations for newly released games. If you want the latest game optimizations for single-GPU, Beta drivers are a must.
 
One of the complaints we receive is we do not test enough games. Now, naturally we can only test what we can test based on time and resources, our method of using real games and testing what is playable takes more time than any other method.

In this review, we take the extra time and include a 6th game, and now make that standard across all reviews, and we still have people complaining. I do hope those that are complaining for whatever reason can see that we are giving it our all to provide the gaming experience in games in each video card review. It is a very intense and timely practice, make no mistake this is not an easy process, but we do listen to our readers and have expanded our suite and will continue to rotate out new games as the months and years go on.

The next game to be added to our suite will be Fallout 4 right around the corner, this means one of the games in the suite will be going bye bye. In early 2016 Deus Ex Mankind Divided is going to be moved in. In this way, we keep evolving and rotating the game suite. I hope this provides some insight into the games we use. If there is a game you don't like to see in the gaming suite, just skip over it, and move on. There will be a time when all games get moved out and replaced after their time with us is done, and as new games came out, we will move new ones in.
 
Without meaning to derail the thread any more, I'd like to know from the people who are claiming bias in game selection (which I call complete horseshit on), what mainstream games that are relevant to a large number of gamers today would you like to see tested?

I see more complaining but still no response to this. It's easy to bitch and moan and be a dick, it's not so easy to actually do something constructive, huh?
 
One of the complaints we receive is we do not test enough games. Now, naturally we can only test what we can test based on time and resources, our method of using real games and testing what is playable takes more time than any other method.

In this review, we take the extra time and include a 6th game, and now make that standard across all reviews, and we still have people complaining. I do hope those that are complaining for whatever reason can see that we are giving it our all to provide the gaming experience in games in each video card review. It is a very intense and timely practice, make no mistake this is not an easy process, but we do listen to our readers and have expanded our suite and will continue to rotate out new games as the months and years go on.

The next game to be added to our suite will be Fallout 4 right around the corner, this means one of the games in the suite will be going bye bye. In early 2016 Deus Ex Mankind Divided is going to be moved in. In this way, we keep evolving and rotating the game suite. I hope this provides some insight into the games we use. If there is a game you don't like to see in the gaming suite, just skip over it, and move on. There will be a time when all games get moved out and replaced after their time with us is done, and as new games came out, we will move new ones in.

You can't please everyone.

You'd have to have 3 versions of each review, and a cookie on everyone's browser based on their forum history to determine which one to show :p
 
100+ people? Really!? Such a sample size. :rolleyes:

Ask yourself this: Are people who own AMD cards playing Project Cars? Are peopl ebuying GPUs playing Project Cars? Is Project Cars one of the newest, top-end (graphically) racing games for PC? Yes? Well then I'd say that it's a viable title to use. Don't you think that a Project Cars player looking to upgrade his card would be interested to see how AMD cards perform? And if it's garbage, is that [H]'s fault or ProjectCars'/AMD's?

Furthermore, think about what you are saying: You want [H] only to test games that perform well on both platforms. WHAT THE HELL GOOD WILL THAT DO?! You'd never test the newest, hardest games (graphically) and all your charts would show equal performance.

Hey man, that's my personal opinion. First time I've heard of Project Cars was from gpu review sites.

I see more complaining but still no response to this. It's easy to bitch and moan and be a dick, it's not so easy to actually do something constructive, huh?

Crysis 3
 
That's it... you're getting the Fillion. ;)
Hey man, that's my personal opinion. First time I've heard of Project Cars was from gpu review sites.
nathan-fillion.gif


Just becuase it's yours doesn't make it right or valuable. Particularly when you throw it at folks with more experience.
86e45efa062fbc5c0c18aa41fb82ff6e.jpg

You mean the game they took off their testing group a while back because it's so old? I mean, it was a great game, and graphically demanding, but it's no longer pushing the boundaries of what's possible.
 
Crysis 3 is the most demanding game I've played on my Ti so far
Moreso than GTA V and Witcher 3 by a huge margin.

I only played the first 2 hours, though.
 
That's it... you're getting the Fillion. ;)

nathan-fillion.gif



You mean the game they took off their testing group a while back because it's so old? I mean, it was a great game, and graphically demanding, but it's no longer pushing the boundaries of what's possible.

Crysis 3 is still more demanding then Project Cars and a lot of other newer games tested. Also please refrain from using Firefly references or gifs here, you are demeaning them with your inappropriate application.
 
Last edited:
Crysis 3 is still more demanding then Project Cars and a lot of other newer games tested. Also please refrain from using Firefly references or gifs here, you are demeaning them with your inappropriate application.

A- Crysis3 may be demanding but it no longer represents the newest in graphical technology. There are new uses of AA, SSAO, textures, tessellation and other tech that it does not employ.
B- it's old and no one (relatively) is playing it anymore.
C- those are both from Castle. Know your Fillion.:rolleyes:
 
I have to say, it's surprising to me how close the 390X performs to the Fury, given the massive shader count boost, HBM, and the optimizations in GCN 1.3 vs 1.1. They must be really handicapping themselves by having not increased the ROP count.

Something is definitely wrong with fiji..............even when you look at the previous drama with AOS. :D
 
If you install new drivers from nVidia or AMD without knowing how they will work in your system and with the games you play, you get what you deserve.

Thing is, that shouldn't be the case.

With nvidia drivers, rarely do new ones cause an issue. It happens but not really that often.

Amd drivers, its 50-50 chance that it will improve performance and fix something, or break something else... it's a bit ridiculous. Users shouldn't have to research every new driver before installing it.

As far as win10 goes, it's running great for me. For the couple friends I know who had some issues with the upgrade such as slow boot times, a fresh install fixed it (suspect some leftover driver from their win7 install, that wasn't fully win10 compatible was causing issues.)

If you did a fresh install and USB doesn't work, that's pretty surprising since Win10 has native usb3 support, win7 doesn't. You shouldn't need to install any drivers for USB3 in win10.
 
A- Crysis3 may be demanding but it no longer represents the newest in graphical technology. There are new uses of AA, SSAO, textures, tessellation and other tech that it does not employ.


Not EVERY game has to be cutting edge in their lineup. People DO play games that are more than 3 months old and Steam stats are not the end all of what people actually play at home.

It is nice to have at least one long term game for consistency sake. [H] has used Tomb Raider over a long period and that was nice having it to compare to older reviews.
 
Okay, variant of GCN 1.2 then if you want to be pedantic :p Fiji has several optimizations that Hawaii does not have.
Is the architecture any different from Tonga?
Fury X seems to be a big fat R9 285 which is probably why it sucks so much.
 
Is the architecture any different from Tonga?
Fury X seems to be a big fat R9 285 which is probably why it sucks so much.

Interestingly, I had a Fury and 2x R9 380 for a while. The shader count is identical between a single Fury and 380 crossfire, and the performance is damn-near spot on the same in games that work well with multi-GPU (synthetic benches like 3Dmark are within 2-3%). I guess they changed some things on Fiji but it really performs like a big Tonga chip rather than Hawaii. Tesselation performance seemed to be up a lot.
 
Interestingly, I had a Fury and 2x R9 380 for a while. The shader count is identical between a single Fury and 380 crossfire, and the performance is damn-near spot on the same in games that work well with multi-GPU (synthetic benches like 3Dmark are within 2-3%). I guess they changed some things on Fiji but it really performs like a big Tonga chip rather than Hawaii. Tesselation performance seemed to be up a lot.
That does make sense, the same tessellation upgrades from Tonga were rolled into Fiji AFAIK.
 
A- Crysis3 may be demanding but it no longer represents the newest in graphical technology. There are new uses of AA, SSAO, textures, tessellation and other tech that it does not employ.
B- it's old and no one (relatively) is playing it anymore.
C- those are both from Castle. Know your Fillion.:rolleyes:
:rolleyes:

How many different ways do people run tessellation?!?! It's still executed the same way on the hardware level with some differences between AMD and NVIDIA....
 
:rolleyes:

How many different ways do people run tessellation?!?! It's still executed the same way on the hardware level with some differences between AMD and NVIDIA....

"New uses of" modified the AA, not the rest of the sentence. The lack of clarity was my fault. :eek:
 
Crysis 3 is still more demanding then Project Cars and a lot of other newer games tested. Also please refrain from using Firefly references or gifs here, you are demeaning them with your inappropriate application.

lLUb55K.gif


This is what you look like to everyone with that silly statement.


Also, on topic, AMD's drivers have always been AMD's downfall. This is coming from a guy who's used ATI/AMD for the last decade....and then I got sick of it and switched to my 980. AMD is going to have to do a lot to get me back.

Makes me sad, but I just don't have the kind of money to waste on simple brand loyalty.

I see very little wrong with [H]'s methods.
 
lLUb55K.gif


This is what you look like to everyone with that silly statement.


Also, on topic, AMD's drivers have always been AMD's downfall. This is coming from a guy who's used ATI/AMD for the last decade....and then I got sick of it and switched to my 980. AMD is going to have to do a lot to get me back.

Makes me sad, but I just don't have the kind of money to waste on simple brand loyalty.

I see very little wrong with [H]'s methods.

This looks better to me.
http://wccftech.com/miscreated-game-cryengines-svoti-fantastic/
 
Not EVERY game has to be cutting edge in their lineup. People DO play games that are more than 3 months old and Steam stats are not the end all of what people actually play at home.

It is nice to have at least one long term game for consistency sake. [H] has used Tomb Raider over a long period and that was nice having it to compare to older reviews.

That doesn't really tell us anything we don't already know. The old game plays great on anything just about, it's not telling us anything about the newer capabilities of GPU's by including it in the game lineup. And the newer capabilities is generally what consumers, smart consumers, want to know about the new GPU's coming out. Newer card = faster than older card. We know this is generally the case. I don't need to see some crysis3 data on my 980Ti to tell me it's faster than my gtx580 or faster than my geforce9800. The speed increases on a really old game such as Crysis 3 will be both in hardware performance and driver improvements. A review from 6 years ago will be using different drivers, whats the point in trying to compare that to today's cards/drivers? They would have to re-test the old card with new drivers (since the new card isn't supported by 6 year old drivers) to get relevant data. The only plus side I can see to doing this would be to see how much a newer driver improved the games' performance vs and old driver. But we already know that driver improvements to any game usually amount to nothing after about 1 year from the game's launch. By then any driver improvements for that game are already done and in the driver.
 
lLUb55K.gif


This is what you look like to everyone with that silly statement.


Also, on topic, AMD's drivers have always been AMD's downfall. This is coming from a guy who's used ATI/AMD for the last decade....and then I got sick of it and switched to my 980. AMD is going to have to do a lot to get me back.

Makes me sad, but I just don't have the kind of money to waste on simple brand loyalty.

I see very little wrong with [H]'s methods.

I don't watch pleb tv anymore so I have no idea what Captain Reynolds is doing after Firefly.

What money were you wasting on brand loyalty? AMD is always within 5-10% of nvidia performance so you are losing what....maybe $50 on FuryX vs 980 TI and that is probably going to change in newer games.
 
The only way that will happen is if AMD releases a new driver that finally turns Fury X into an "overclocker's dream". It's lagging by 20-30% right now... You're a very optimistic person.
 
Depends on the resolution. Highly unlikely the overclocking will improve, but AMD does tend to improve the performance relatively significantly with drivers over the next year or so. That said, at the same price point ($649) the 980 Ti is the better card.
 
And the newer capabilities is generally what consumers, smart consumers, want to know about the new GPU's coming out. Newer card = faster than older card. We know this is generally the case. I don't need to see some crysis3 data on my 980Ti to tell me it's faster than my gtx580 or faster than my geforce9800. The speed increases on a really old game such as Crysis 3 will be both in hardware performance and driver improvements. A review from 6 years ago will be using different drivers, whats the point in trying to compare that to today's cards/drivers? They would have to re-test the old card with new drivers (since the new card isn't supported by 6 year old drivers) to get relevant data. The only plus side I can see to doing this would be to see how much a newer driver improved the games' performance vs and old driver. But we already know that driver improvements to any game usually amount to nothing after about 1 year from the game's launch. By then any driver improvements for that game are already done and in the driver.

I never implied that they keep a game around for 6 years Mr. Smart Consumer. Two to three years would be nice, especially since the small improvements have been made in 3 years, ie stuck at 28nm, 7970 vs. 280x vs. 380x. It would be nice to see what is worth adding a used card and what is worth upgrading. As for the drivers, I couldn't follow you since you contradict yourself saying that drivers matter, then not so much or whatever.
 
I don't watch pleb tv anymore so I have no idea what Captain Reynolds is doing after Firefly.

What money were you wasting on brand loyalty? AMD is always within 5-10% of nvidia performance so you are losing what....maybe $50 on FuryX vs 980 TI and that is probably going to change in newer games.

Look everyone, he doesn't watch TV anymore!

See, nobody cares. Stop bringing it up.

There's a lot more to a user experience than 5 - 10 % performance lead. Driver stability, power draw, cooling, sound, etc. The whole package. AMD can engineer a fine product, but their drivers are something that I'm sick of working with. I think the 290X was where they started to go wrong with this philosophy. I would know, I had a 290X for almost a year. I had to water cool it to maximize its performance, and it was an absolute space heater. Once I replaced it with my current 980 GTX, I didn't just get a performance gain, I was able to have it without having a GPU loop, with lower energy cost. And drivers that work, without me having to constantly install a beta driver. The only thing I miss is Mantle support. But that's really not turning into the thing it was pimped out to be.
 
Look everyone, he doesn't watch TV anymore!

See, nobody cares. Stop bringing it up.

Have you heard of HBO or Netflix? Sure they cost a bit extra but very worth it.

There's a lot more to a user experience than 5 - 10 % performance lead. Driver stability, power draw, cooling, sound, etc. The whole package. AMD can engineer a fine product, but their drivers are something that I'm sick of working with. I think the 290X was where they started to go wrong with this philosophy. I would know, I had a 290X for almost a year. I had to water cool it to maximize its performance, and it was an absolute space heater. Once I replaced it with my current 980 GTX, I didn't just get a performance gain, I was able to have it without having a GPU loop, with lower energy cost. And drivers that work, without me having to constantly install a beta driver. The only thing I miss is Mantle support. But that's really not turning into the thing it was pimped out to be.

Now I know you are trolling. AMD never had drivers burn up graphics cards.
 
Have you heard of HBO or Netflix? Sure they cost a bit extra but very worth it.



Now I know you are trolling. AMD never had drivers burn up graphics cards.

Seriously? When did I say anything about drivers burning up graphics cards? I'll wait for you to quote me.
 
Seriously? When did I say anything about drivers burning up graphics cards? I'll wait for you to quote me.
There's a lot more to a user experience than 5 - 10 % performance lead. Driver stability, power draw, cooling, sound, etc. The whole package. AMD can engineer a fine product, but their drivers are something that I'm sick of working with. I think the 290X was where they started to go wrong with this philosophy. I would know, I had a 290X for almost a year. I had to water cool it to maximize its performance, and it was an absolute space heater. Once I replaced it with my current 980 GTX, I didn't just get a performance gain, I was able to have it without having a GPU loop, with lower energy cost. And drivers that work, without me having to constantly install a beta driver. The only thing I miss is Mantle support. But that's really not turning into the thing it was pimped out to be.

Did you even notice what I bolded? Its like Pot calling silverware black; trying to claim NVIDIA drivers are more stable and better when last time I checked it was NVIDIA drivers killing NVIDIA cards. Can't remember when ATI pulled that one off.

Since NVIDIA drivers control more and more of the power delivery systems on their GPUs you better pray that NVIDIA driver team doesn't fuck up again and send a precious TitanX to gpu heaven.

http://modcrash.com/nvidia-display-driver-damaging-gpus/#.UbS1ePnvtQs
 
Last edited:
I have a Nvidia Geforce GT 650M in my notebook and two Radeon 280xs in my desktop. Nvidia's drivers are actually more unstable for me most of the time and more likely to glitch out. I'd had better results with the AMD drivers, with the exception of Crossfire which either doesn't work or is slower in a few games. Namely Guildwars 2 (Still not fixed, despite claims to the contrary), Titanfall (Annoying texture flickering) and Sky Forge (unplayable with crossfire on). From memory, I seem to remember the experience being slightly better with my previous Geforce 460s in SLI.
 
Seriously? When did I say anything about drivers burning up graphics cards? I'll wait for you to quote me.

I am not sure why he quote that, no where did you mention drivers are causing heat issues to your 290X. I mean the 290X draws more power which means will gives off more heat than a 980, test done by [H] shows that, it is like he takes two unrelated statements regarding your driver issues and heat and made it into a related statement. As for AMD drivers, never own a 290X so I cannot comment what driver issues you have, but I never had issues with either camp drivers.
 
Nice article. To me it looks like the real winner is the 390x, cheapest of the bunch, and typically pretty close performance to the R9 Fury, while still beating the 980's most of the time!

Sure do suck gobs of power but hey, if that's the only downside, it wouldn't stop me.
 
Back
Top