[computerbase] Project CARS benchmarks

[exar333]

I find it really hard to believe the Nvidia driver team does not have the resources to continue optimizing for Kepler.

In some ways, it reminds me of the Windows Vista and Win7 driver 'issues' from HP a few years back. They just didn't write drivers for the 'older' (read: all but current models) printers and EOL'd them essentially forcing users to buy new printers if they wanted the new OS.

Its not the same thing, but forced-obsoleteness by lacking to continue innovating and just 'keeping the lights on' for Kepler isn't too dissimilar.

By no way do I know for sure NV is failing to enhance Kepler, they aren't purposefully handicapping it, but it doesn't look like much effort is being put into optimizing for it. I guess that's theur choice. If people still buy their GPUs regardless, then I guess I cannot blame them.

Personally, I greatly appreciated the length of time my 5870 got me through gaming. My 670 didn't last that long, and my 970 definitely will not either, but part of that is just the 28nm mess we are in...

 

[RussianSensation]

Kepler sure did take a dive in this title -- that's an eye opener. Thanks for the link.

Even if we ignore AMD's and Kepler's performance though, check out that chart:

970 = 46 fps
980 = 47 fps
Titan X = 48 fps

The game's engine seems to have major CPU (other?) bottlenecks. This type of scaling is simply unacceptable. It's possible there are certain parts of this game engine going back to NFS Shift series that are hogging down performance on all GPUs.

Also, if you look at AMD's or NV's performance from 1080P to 4K, the performance drop off is too small to make logical sense (4x the pixels but performance often drops off just 15-31% for NV cards).

1080P, rain
Titan X = 48 fps
GTX980 = 47 fps
780Ti = 39 fps

4K, rain
Titan X = 41 fps (-15% drop)
GTX980 = 33 fps (-30% drop)
780Ti = 27 fps (-31% drop)

I am not a programmer but it appears there are some major CPU bottlenecks or game engine issues at lower resolutions in this game. There is no way going from 1080P to 4K should only be a 15-31% drop in FPS. To me it doesn't appear that this game is well optimized to start with.

I would like to see benchmarks of this titles after various developer patches, NV/AMD driver updates and all done on Windows 10 where CPU overhead is reduced. I bet performance will improve for everyone.

I believe there is a lot more to it than what is being repeated on the forums. Kepler performance has been hurting in a lot of recent titles, that is for sure.

Of course AMDs GCN isn't showing its age, it is still their current line up. It is AMDs current gen. On to of that, GCN is now in the current gen consoles. That was supposed to give them an edge and I think people just ignore that. Then, lets not forget that AMD had some pretty notable big driver improvements over the years.

Those are just a few things that any level headed person could easily see. Then there is the flip side. If people comparing recent performance of Hawaii and Tahiti to Kepler just put it off as nvidia abandoning Kepler, they the doing a disservice to AMD. AMD put a lot of effort in getting the most out of the gpus and it is finally paying off well.

I think Tahiti was a very beefy design. It had some killer specs. So much so that on paper there is no way the gk104 should have been able to touch it. It makes so much more sense to me that AMD just got better at harnessing that power.

I remember I did a mathematical calculation and I compared Kepler to Maxwell, Kepler to GCN and GCN to Maxwell and the result was that GCN narrowed its gap to Maxwell by a couple percent but in the same time Kepler fell far behind both GCN and Maxwell in the same period. It's hard to believe that almost immediately after GTX970/980 launched that by pure chance Kepler started bombing and especially so in GW titles.

For example, Computerbase has 960 just 6% within a 780Ti in their testing. NV has publicly stated that Maxwell has about a 35-40% increase in IPC per CUDA core (128 Maxwell cores are 90% as fast as 192 Kepler cores). It's impossible to explain how a 1024 CUDA core 960 can be so close in performance to 2880 CUDA core 780Ti unless NV improved IPC by 2.5X. There are so many titles now where the 960 is near 680's level of performance, 970 is beating 780Ti, 980 is far ahead of 780TI sometimes by 25-35%, 780 barely beating 7970Ghz!

Tessellation can't be the reason because Maxwell hardly improved it over Kepler. Memory bandwidth can't be the reason because 780Ti ha at least as much as a 970 even if we account for 33% more efficient use of memory bandwidth by the 970. Just look at the number of texture mapping units of 780Ti vs. 970. Unless all Kepler cards are suddenly pixel fill-rate limited, what is the explanation that driver focus by NV is now on Maxwell? Also, if pixel fill-rate was an issue, 7970Ghz would start taking even more but its lead over 680/770 keeps growing and its performance is barely behind a 780.

Some have said most of modern titles utilize a lot more compute shaders in the pipeline but I have yet to see any developer claim they are using a lot of DirectCompute in their games, especially not in GW games because that would instantly give GCN a major advantage even over Maxwell.

We probably won't ever know if NV stopped optimizing drivers for Kepler and refocused all its energy on Maxwell but for a consumer who bought $650 780, $1K Titan, $700 780Ti, it doesn't particular matter what the reason is -- when they launch the game and see a $180 960 nearly as fast as their $700 780Ti or the ancient 7970Ghz being a hair behind 780 and R9 290X now matching or beating 780Ti, I doubt they are happy about this. I don't recall GTX480/580 falling apart this badly just 1.5 years from their launch. GTX580 performed about as well as a 660Ti/7870 and started to suffer a lot later on due to 1.5GB VRAM.

That's all fair enough, ocre, and it'd be hard to tell some of that.

However there should probably be a thread comparing:

1. GCN and Kepler once both had been launched
2. GCN, Kepler and Maxwell at Maxwell launch
3. GCN, Kepler and Maxwell now

All on consistent games released before that point. So 1 gets a few games, then 2 adds a few more and gives a second look at later performance in game set 1 and so on.

It would definitely be interesting to see how they fared next to each other, and whether for example, Kepler has gotten left behind without improvements or Maxwell has improved significantly more than Kepler.

Someone did a very extensive analysis of this and it does show Kepler performance fell off a cliff following 970/980's launch. For 12 months leading up to 970/980 launch, 780/780Ti performed well against 7970Ghz/R9 290/290X.

https://www.reddit.com/r/buildapc/comments/2one2z/discussion_has_nvidia_forsaken_kepler_cards_has/

"General conclusions

First of all, the relative performance of Nvidia cards compared to AMD cards, for both Maxwell and Kepler, is down across the board. For example, at 1080p, the 970 which previously had an 8.0% lead over the 290X now finds itself only 4.1% ahead. The Titan, which had a 2.3% performance advantage over the 290X, now finds itself trailing the AMD card by 4.1%. At 4K, the story is similar: the 980’s 11.2% performance lead over the 290X is down to 8.6% and the 780 Ti, which used to beat the 290 X by 3.1%, is now losing by 1%. However, another conclusion that can be drawn is that the relative performance of Kepler cards has dropped by much more than their Maxwell counterparts’. At 1080p, while the 980 goes from beating the 290X by 22.7% to having a 17.3% lead, the 780 Ti’s 14.7% lead over the 290X drops to a lowly 6.1%. This can be seen by comparing the two Nvidia cards, and the 980’s 6.9% performance advantage over the 780 Ti grows to 10.6%. And at 4K, while the 970’s deficit over the 290X grows from 3.1% to 4.8%, the Titan’s 5.1% deficit turns into a huge 11.4% loss to AMD’s card. The 970’s relative performance thus grows from beating the Titan by 3.3% to 7.5%. These numbers might not seem significant, but bear in mind that they are caused by a handful of new games."

Take a look at Sweclockers

Sept 19, 2014 - Maxwell launch

980 is 16% faster than a 290X (Uber)
780Ti is 9% faster than a 290X (Uber)
Titan is 29% faster than 280X

vs.

980 is 11% faster than 290X (Uber)
780Ti is tied with 290X (Uber)
Titan is 20% faster than a 280X

Kepler lost way more performance than Maxwell vs. GCN since Sept 2014. Whether Kepler architecture is becoming outdated, or NV is not focusing on the drivers as much -- someone who bought last gen's $600+ Kepler cards can't be happy with this!

That's why I keep advocating for PC gamers to upgrade more often than trying to buy $600-700 cards to future-proof for 3-4 years. You never know if AMD's or NV's architecture will get crippled by drivers/bottlenecks in 2 years. Historically though, if we go back to HD4890 vs. GTX280 and until today's line-ups, AMD cards tend to perform better in more demanding next gen titles as they age, while NV's advantage more or less disappears from 15-20% to single digits (think 6-7%) as time goes on. The one big exception to this rule is the HD5850/5870 series which got killed against 470/480/580 cards due to poor tessellation and having only 1GB of VRAM.

 

[loccothan]

RussianSensation + 1 My Man for the eye opener....

-> Here intresting Mod i have for p^Cars... (it's my Idea generally)
Just imagine how much FPS you can gain after switch some files with AC:BF and FC4 files. Especially for Particles/rain effects.
Maby this will help nV brothers with Older GPU's like 6xx/7xx

-> http://forums.anandtech.com/showthread.php?t=2430596

 

[Enigmoid]

"General conclusions

First of all, the relative performance of Nvidia cards compared to AMD cards, for both Maxwell and Kepler, is down across the board. For example, at 1080p, the 970 which previously had an 8.0% lead over the 290X now finds itself only 4.1% ahead. The Titan, which had a 2.3% performance advantage over the 290X, now finds itself trailing the AMD card by 4.1%. At 4K, the story is similar: the 980’s 11.2% performance lead over the 290X is down to 8.6% and the 780 Ti, which used to beat the 290 X by 3.1%, is now losing by 1%. However, another conclusion that can be drawn is that the relative performance of Kepler cards has dropped by much more than their Maxwell counterparts’. At 1080p, while the 980 goes from beating the 290X by 22.7% to having a 17.3% lead, the 780 Ti’s 14.7% lead over the 290X drops to a lowly 6.1%. This can be seen by comparing the two Nvidia cards, and the 980’s 6.9% performance advantage over the 780 Ti grows to 10.6%. And at 4K, while the 970’s deficit over the 290X grows from 3.1% to 4.8%, the Titan’s 5.1% deficit turns into a huge 11.4% loss to AMD’s card. The 970’s relative performance thus grows from beating the Titan by 3.3% to 7.5%. These numbers might not seem significant, but bear in mind that they are caused by a handful of new games."

While Nvidia stopping optimization for Kepler is one explanation, another is that games simply became more optimized for GCN (and AMD released some performance drivers) and started using more compute. The fact that AMD gained ground over maxwell supports the idea of more optimization over maxwell as Nvidia has been consistent with their driver releases and has no reason not to optimize for Maxwell.

http://techreport.com/review/27969/nvidia-geforce-gtx-titan-x-graphics-card-reviewed/3

In terms of synthetics, the 780 Ti had no business competing with Hawaii let alone beating it.

 

[Eymar]

Even if we ignore AMD's and Kepler's performance though, check out that chart:

970 = 46 fps
980 = 47 fps
Titan X = 48 fps

The game's engine seems to have major CPU (other?) bottlenecks. This type of scaling is simply unacceptable. It's possible there are certain parts of this game engine going back to NFS Shift series that are hogging down performance on all GPUs.

I have to agree this game (while very good graphics wise) has a CPU bottleneck that probably can be optimized. I'm running a 4.4Ghz 5930k with SLI Titans and one core is always pegged at 100% or close to it during games. If I set power options to max process state to 80% (which maxes CPU hz to around 2.9Ghz, have to double check) the FPS goes down proportionally (ex. 70FPS to 44FPS). Another weird behavior is if I limit max processor state to 80% the cpu usage on the one core that goes to 100% to 60%.

 

[RussianSensation]

In terms of synthetics, the 780 Ti had no business competing with Hawaii let alone beating it.

How can you make such a sweeping statement? First of all, NV has had a superior architecture for tessellation. That means by all accounts 780Ti should and did beat 290X in that metric. Secondly, NV manufactured a 561mm2 die and Hawaii was a 438mm2 die. Your statement implies that we should automatically expect a 438mm2 Hawaii to beat a 561mm2 NV card in synthetics?

The compute argument doesn't work in Project CARS because GCN is amazing at compute but it's beaten by a 960. Also, the compute argument doesn't explain how the Titan is barely faster than a 960, or how 970/980/Titan X hardly scale in the rain at 1080P. If Compute was the limiting factor, we should see linear scaling between 970 to 980 to Titan X in Project CARS. By similar account, we would never see just a 15-16% scaling from a 280X to 290X if compute was a key factor. That's not what happens in this title.

As far as other games go, your "Compute is the answer" position does not at all address how Kepler's performance is so good in almost all vendor agnostic titles released in the last 8 months. Are you suggesting only GW's and AMD GE titles are Compute heavy but most other PC games do not use compute? GameGPU.ru shows plenty of titles where 780TI > 970 and is even beating 290X in non-GW titles. To suggest that compute is the primary factor that explains why Kepler's performance fell apart is to insinuate that all other games where 780/780TI/Titan perform well have little to no compute. Do you have direct evidence of that?

 

[loccothan]

But maby, just Big Maby they intentinally cripple PC version of pCars to boost sales for PS4/X1 ? So, the consoles games aren't cheaper right?
How's that
And another proof for my Maby is that SMS working Hard on Optimisations but ! for PS4 ;-) he he... You'' be waiting and waiting.......

But there's another realy Great SIM game for You out there -> Assetto Corsa, great physics, real roads, Mods, optimised very Good and only PC ! What you need more, some Arcade/sim opimisation crap ;-)
There's always a choice. Greets

 

[ocre]

RS, I have no idea how you can manipulate the situation to such a degree. Your explanation of what is going falls way short.

In case it really has escaped you, which I find hard to believe, let's look at what it means.

Using the exact same data you posted, of course. See, the only way your scenario is correct, we all have to accept that nvidia simply could not improve maxwell at all. That for the first time, we have a new architecture that doesn't gain performance over time. I don't know how many times you posted Tahiti performance over the 6970 at launch to its performance 1yr + down the road.
So in that case, the gains AMD made were just a result of them abandoning the 6970? I have never heard you spin it like that. As a matter of fact, never once have you in the countless times you brought up driver improvements over time. Just now, all of a sudden you flip everything upside down with maxwell.

So lets look at the figures and charts you posted. Just that example.

It is clear that the 290x has gained ground against the 980. The 16% gap dropped to merely an 11%. That is 5%.

So when the 780ti was only 9% faster at the start, if all things were equal, then there is 4% lost. That is the difference. The 9% advantage should have dropped to 4% if AMD had only gained 5% with GCN improvements over this time.

Is it really that hard to imagine that nvidia improves maxwell 4% with more mature drivers? In that case, AMD would have actually increased GCN performance, not the 5% we see when comparing it to the 980, but that AMD actually was able to extract 9% more performance. Strangely enough (not), that fits much better with the omega driver update. The omega driver update that AMD released graphs showing improved performance of 9 to 19% for the 290x. Feel free to look it up. That all by itself is enough to close the gap between the 290x and the 780ti. Remember, it was only 9% faster than the 290x when the gm204 launched.

There really isn't any room for wild conspiracy theories. AMD realized a rather potent driver update and you are completely throwing that out the window.

I propose that this driver update improved the 290x and GCN performance, AMD tells us they have, we actually have reviews showing they have, all signs point to AMD improving GCN performance.

So instead of ignoring this, lets look at the data you posted again. AMD I proved 290x performance 9-20%, this was enough to close the gap between the 780ti and the 290x. Nvidia working with the fresh new maxwell architecture was able to bring performance up a little as well. All they need is 4% and everything adds up perfectly.

So do you really expect everyone to believe that nvidia could not improve maxwell at least 4% since it launched? Do you believe this? What about AMD, you really don't think they made about a 10% jump in performance with the omega driver?

The truth is, AMD had untapped potential. The truth is, nvidia extracted about all they could out of Kepler. This is obvious when you look at never settle, the time the 7970ghz pulled an undeniable lead over the gtx680. For many many months, nvidia lost the performance crown. This is because, nvidia couldn't respond, they already were extracting about all they could out of the 680.

If nvidia couldn't gain back the lead when the 7970ghz topped the 680 with driver improvements, why now when the 290x topped the 780ti do you have to invent a whole new conspiracy?

With the same data you just posted, clearly there is a more reasonable factor at play here. The only way your theory is true, we all have to say that Nvidia just couldn't extract performance from maxwell and we have to throw out all the improvements AMD has made as well. Just take away all the credit that AMD is completely due. Its not AMD doing better, it must be nvidia doing worse.
Well. I can't change your mind, but I am sure others will read this and see that your theory is not the answer.

There is way more at play here, obviously.

 

[3DVagabond]

*snip*

not disagreeing with anything you said. There's logic there. I just want to say that this completely flies in the face that AMD DX11 drivers are so bad compared to nVidia and they don't have resources, are just concentrating on DX12, and so on.

So, next time someone uses all of those reasons to explain why AMD doesn't do so well in some games, like the game in this thread, please remember everything you just posted.

 

[raghu78]

Project Cars has some serious bottlenecks which is why we do not see the Titan-X 30 - 35% faster at 1440p and 35-40% faster at 4k than GTX 980. Titan-X has 50% more resources than GTX 980 in every aspect - cc, tmu, rop, bandwidth. Even with the slightly lower clocks wrt GTX 680 we see roughly 35% gain consistently in most games. So definitely the game needs patches to improve performance.

http://www.computerbase.de/2015-05/...diagramm-grafikkarten-benchmarks-in-2560-1440

The perf scaling from 980 to titan-x at 1440p is just 13% and at 4k is 21%. thats really bad.

As for AMD cards they need a driver update and maybe also some game patch update to fix performance. I have to say AMD have been pretty slack in this title. This kind of perf issue must have been resolved before launch.

 

[3DVagabond]

As for AMD cards they need a driver update and maybe also some game patch update to fix performance. I have to say AMD have been pretty slack in this title. This kind of perf issue must have been resolved before launch.

Well considering Ian Bell said they hadn't had any contact since something like October last year, and then after looking into it realized they had many times been contacted by AMD, it's hard to say who, if anyone, was slack.

 

[raghu78]

Well considering Ian Bell said they hadn't had any contact since something like October last year, and then after looking into it realized they had many times been contacted by AMD, it's hard to say who, if anyone, was slack.

This problem with Gameworks licensees is getting irritating and AMD must go public on record with the tech press saying that Nvidia is harming them by influencing Gameworks licensees to drag their feet with AMD GPU performance.

Why the heck should it take Ubisoft 4 months from launch for a game update to enable CF or Techland roughly 6 weeks from launch to fix performance in AMD GPUs through a game patch.

 

[3DVagabond]

This problem with Gameworks licensees is getting irritating and AMD must go public on record with the tech press saying that Nvidia is harming them by influencing Gameworks licensees to drag their feet with AMD GPU performance.

Why the heck should it take Ubisoft 4 months from launch for a game update to enable CF or Techland roughly 6 weeks from launch to fix performance in AMD GPUs through a game patch.

I talked to an admin (who I won't throw under the bus) for a tech site recently asking him why "certain things I see in reviews happen". He said nVidia have very strict guidelines if you want to receive the latest and greatest review gear from them (He said AMD does also. Just not as strict) and get invited to their events all around the world you follow them. AMD doesn't have the same kind of budget for that stuff. nVidia invites more people to more events and distributes more product for review.

Go back and look at how many 980/970 reviews there were when those cards were released. Then compare when AMD releases a card.

 

[raghu78]

I talked to an admin (who I won't throw under the bus) for a tech site recently asking him why "certain things I see in reviews happen". He said nVidia have very strict guidelines if you want to receive the latest and greatest review gear from them (He said AMD does also. Just not as strict) and get invited to their events all around the world you follow them. AMD doesn't have the same kind of budget for that stuff. nVidia invites more people to more events and distributes more product for review.

Go back and look at how many 980/970 reviews there were when those cards were released. Then compare when AMD releases a card.

I understand. Money talks and the average human mind (in the tech press) is just too weak to stand for ideals like integrity, fairness and ethics.

 

[Skurge]

Well considering Ian Bell said they hadn't had any contact since something like October last year, and then after looking into it realized they had many times been contacted by AMD, it's hard to say who, if anyone, was slack.

They did? well that is lame from SMS.

 

[SirPauly]

This problem with Gameworks licensees is getting irritating and AMD must go public on record with the tech press saying that Nvidia is harming them by influencing Gameworks licensees to drag their feet with AMD GPU performance.

I'm all for this: If Gameworks is actually doing harm --- let AMD prove its case -- or third party investigations prove it also --- much better than wild conspiracy theories that claimed GPU physics when there wasn't any.

 

[Makaveli]

I understand. Money talks and the average human mind (in the tech press) is just too weak to stand for ideals like integrity, fairness and ethics.

This reminded me of a quote from one of my bosses many years ago.

"Integrity,fairness and ethics don't pay the bills!"

 

[Silverforce11]

I'm all for this: If Gameworks is actually doing harm --- let AMD prove its case -- or third party investigations prove it also --- much better than wild conspiracy theories that claimed GPU physics when there wasn't any.

 

[desprado]

He is talking about Gamework Features not the whole game just like TressFx 2.0 in Lichdom.

 

[Erenhardt]

"Cannot Be Optimized".... "Cannot Be Optimized".... what?
I guess there is no technical reason, just their obligation.

I feel encouraged to not pay them if the performance is below my expectations...

Project cars is the same, but worst, as you can't disable those options.

 

[ocre]

not disagreeing with anything you said. There's logic there. I just want to say that this completely flies in the face that AMD DX11 drivers are so bad compared to nVidia and they don't have resources, are just concentrating on DX12, and so on.

So, next time someone uses all of those reasons to explain why AMD doesn't do so well in some games, like the game in this thread, please remember everything you just posted.

I have been saying this for some time now. I am usually so short on time that I don't get to include raw data. RS posted the raw data (which i do appreciated), I just applied what I have been suggesting and it lines up just like you would expect.

AMD has had some really large performance jumps since GCN but I think the fact that they still were extracting perfrormance jumps like they have with omega, that is amassing.

I just hate that people would rather ignore the work and progress AMD has made just to try to make nvidia look bad. It is a disservice to AMD if you ask me. Not only has AMD surpassed the 780ti with Hawaii, they have made a rather impressive jump on the 980. That is nothing short of amassing.

Now, , I do believe there are other factors at play here. I believe that GCN in the consoles has been beneficial, which at one time I didn't see how it would help them on PC .

It is not so strange to me when I see the GCN besting Kepler like it is. Spec wise, AMD cards were more beefy. Gflops, told the story if the paper specs were not enough. The gk104 actually keeping up with Tahiti, if we look at the specs of both designs, that is what is hard to imagine. Nvidia pulled all the tricks and clever programming they could, but ultimately AMD leaps over the gk104 with Tahiti.

Then we have Hawaii vs the gk110. Some people may want to focus on die sizes, a completely useless measure. The truth is that Hawaii had a much greater transistor density. But, even that doesn't mean much. It is not the transistors that count, it is how they are used. The design. How many of them are even used when it comes to gaming.
Hawaii, it had some very impressive and beefy specs, on paper and according to their gflops. The 290x best the 780ti in gflops, so in a perfect world, as long as the cores can stay completely fed and saturated, it can be more powerful.

It is hard to keep the entire GPU fed and all parts of the pipeline saturated. It takes a lot of work and talent. In theory the 290x can be more powerful than the 780ti, at least when we look strictly at gflops. So should it be so surprising that AMD has recently found better ways to extract that power.
It is not so surprising to me. It actually makes perfect sense.

 

[Headfoot]

I wouldn't doubt CPU bottleneck at all. They crank these games out at an alarming rate. They seem to add nice new visual effects in each iteration in addition to plenty of content (cars, tracks, UI, etc.). Either they have a massive team, or something is getting left on the table. They're probably just frankensteining an old engine that isn't super well threaded. And I can see why -- much easier to make money when you can drop a game each year and only have to pay for developing enhancements instead of rewrites.

At a relatively early point, you can't optimize CPU without re-architecting from the ground up for native multithreading. There is only so much optimization you can write in after the original architecture is designed wrt CPU on DX11

 

[SirPauly]

That's too bad, and sad to see. Without AMD not able to optimize may hurt their customers even though the feature is brand agnostic.

 

[3DVagabond]

That's too bad, and sad to see. Without AMD not able to optimize may hurt their customers even though the feature is brand agnostic.

They are not just AMD's customers. They are CD Projekt Red customers too. Why would anyone want to do business with a company that will treat their customers this way? I would recommend staying as far as possible away for everyone. This time it's nVidia that paid the protection money, next time it might be AMD? Then all the people who are saying not my problem now will be singing a different tune.

 

[desprado]

They are not just AMD's customers. They are CD Projekt Red customers too. Why would anyone want to do business with a company that will treat their customers this way? I would recommend staying as far as possible away for everyone. This time it's nVidia that paid the protection money, next time it might be AMD? Then all the people who are saying not my problem now will be singing a different tune.

AMD is not in finical position to pay for 74% DGPUs user protection money to developers.Majority always speaks and with 20% share left with AMD put them in a very tight position .