[fixed] Dragon Age 2 Low Performance on Nvidia Cards [fixed]

[Keysplayr]

So There's a problem with him not waiting? He bought the game, he bought the cards. He has every right to be pissed.

What does that have to do with him waiting for a fix? Pissed or not?

Wouldn't you wait until a fix was available for a much better gaming experience? Or would you plod through it when something isn't right?

 

[Silverforce11]

The fact that its a problem in dx10 mode as well for NV cards suggest this is a bigger problem and may have to do with the game engine itself.

 

[tviceman]

So There's a problem with him not waiting? He bought the game, he bought the cards. He has every right to be pissed.

I never said he didn't have a right to be pissed, nor did I imply that. I meant that by playing through the game now, he'll just complain more when a performance fix is released after he's beaten it or played through most of it.

 

[CVSiN]

Sure the game doesn't need 60 fps to be playable. But let's take a closer look to the situation shall we?

You have a gtx 580 - the fastest single-gpu on the market from nvidia. Not everyone can afford that, not even gonna go into sli-discussions. Yet not even on that card does it run fully smooth (60 fps). It's more around 40-45.

Personally i have a gtx 560 ti and average around 30-35 fps. I don't find that performance acceptable. 460s, 470s and 480s have even worse performance. You see, this game should be running on very high, maxed out on mid-range cards at 45~ish fps. High-end cards should max it np with constant 60+. Just like ATIs.

Simply because you don't really care to even check your fps it shows that you'd consider anything "smooth" as long as it doesn't have stutters (that includes 30 fps which is in no way smooth).

http://translate.google.hr/translate?js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&sl=auto&tl=en&u=http%3A%2F%2Fwww.pcgameshardware.de%2Faid%2C815403%2FDragon-Age-2-Grafikkarten-Benchmarks-mit-DirectX-11-und-DirectX-10-Radeons-vor-Geforce%2FRollenspiel-Adventure%2FTest%2F

That's a link to german benchmark of DA2. All Nvidia cards are getting crushed by ATI. They manage to perform twice as fast. Might not affect you, but it sure annoys me when i payed more for this card than i would have for an ati and at the same time get half the performance.

Ok fine you wanted numbers.. according to EVGA Precision outputting FPS on my G15 keyboard its never dropping below 50 FPS while running around or fighting..
in talking scenes 70+ fps

this is with everything set to flat max. DX11 with texture pack at 1920x1080 in full screen mode

I cannot get it to lag or stutter 50 FPS is smooth to my eyes..

I still really dont see the issue.. it's performing just fine.
if you have lower end cards lower your settings just as in other games..

I also have a watercooled i7 EVGA GTX 460 SLI system as well but have not run DA2 on it yet as it needs a new X58 MB as the Sabertooth died on it. Would love to test that as well..

 

[SHAQ]

Nvidia had 6 months to optimize for the game. Its not like it came out of nowhere. I'm not too angry because I don't think the game looks much better on very high anyway. I can see the game dropping frames where it shouldn't and even vsync on lower settings has stutters in it. They had plenty of time. If Crysis 2 runs 2x better than DA2 on Nvidia then it might be okay but I'm sure it won't. They did pay them $2mil as I understand. I don't know if AMD payed EA/Bioware anything.

 

[VirtualLarry]

Wow, what a pickle NV seems to find themselves in. Glad I went with CF HD6870s for my BILs gaming rig. He probably picked up this game by now. Hope he upgraded the drivers.

 

[Skurge]

I never said he didn't have a right to be pissed, nor did I imply that. I meant that by playing through the game now, he'll just complain more when a performance fix is released after he's beaten it or played through most of it.

OK, now I understand what you meant

 

[chris2012]

I'm not really playing it atm. Gonna wait for a fix instead but it's rather sad that i have to do that on an AAA title that's all.

And to CVSIN yes 50 fps is bad on a 580. You should have 60 constant. I should have 50 on my gtx 560ti. And no it's not me having retarded expectations. As long as a 6950 can pull 45-50 fps so should I. Hope you get the point.

The only option you can lower to increase performance is going from dx11 effects down to 10 by using high option (you lose tesselation, advanced lighting, blur, pp dof etc). It makes no difference whatsoever if i disable dof/blur/aa/af while using very high. That's one of the issues in the first place. Bad coding and bad driver, hello.

 

[Grooveriding]

I'm not really playing it atm. Gonna wait for a fix instead but it's rather sad that i have to do that on an AAA title that's all.

And to CVSIN yes 50 fps is bad on a 580. You should have 60 constant. I should have 50 on my gtx 560ti. And no it's not me having retarded expectations. As long as a 6950 can pull 45-50 fps so should I. Hope you get the point.

The only option you can lower to increase performance is going from dx11 effects down to 10 by using high option (you lose tesselation, advanced lighting, blur, pp dof etc). It makes no difference whatsoever if i disable dof/blur/aa/af while using very high. That's one of the issues in the first place. Bad coding and bad driver, hello.

I was looking forward to the game too much to wait. It's unfortunate nvidia hardware is so poor in DX11 in this game.

Some thoughts on the Nvidia forums are that since the game is using HDAO under DX11 this the problem due to AMD being superior in HDAO and nv performing very poorly.

 

[Seero]

Ok fine you wanted numbers.. according to EVGA Precision outputting FPS on my G15 keyboard its never dropping below 50 FPS while running around or fighting..
in talking scenes 70+ fps

this is with everything set to flat max. DX11 with texture pack at 1920x1080 in full screen mode

I cannot get it to lag or stutter 50 FPS is smooth to my eyes..

I still really dont see the issue.. it's performing just fine.
if you have lower end cards lower your settings just as in other games..

I also have a watercooled i7 EVGA GTX 460 SLI system as well but have not run DA2 on it yet as it needs a new X58 MB as the Sabertooth died on it. Would love to test that as well..

To some, it is a problem. Forget about AMD, run the game in Dx9 and check FPS. It is far greater than the FPS you are getting in Dx11. Now is it tessellation done right? or was there something wrong? Maybe there are some things that are not enabled on AMD cards, or maybe there exists some pieces of code that cause extra loops on Nvidia cards, or even errors that are not visible to users.

Again, smooth or not is subjective. Some only sees the number of frame per second shown by whatever program they use to produce that with. Some, like you, don't give a @$#% as long as it is smooth. Note that your PC is top of the line, but many does not have your setup. The game really shouldn't require 580 to produce a reasonable frame rate with everything maxed. A user can choose to be forgiving on performance, and some can seek for solutions.

If seeking solution is the goal, or not, trying out loud, spreading rages, or even start to create self inflict wounds won't solve the problem. If the problem indeed resize on the driver, waiting for another driver update is the only solution. If it is game code related, then there can be a few solutions.

Bottom line, Dx11 isn't something new anymore. There were many games out and none of them cause problems like that. IMO this kind of problems should have not existed for a released game. It harms the game more than anything else. Why this is happening isn't our primary subject to worry. Getting a solution is. Well, after playing the game of course. 50 FPS is ok for me too, but i will like to play it in 3D, with probably ends up being 25 FPS, not tolerable. That may be the reason why some people are upset about it.

 

[Qbah]

CVSiN is getting 50FPS, yet everyone else, even people having GTX580 SLi (or GTX480x3!) are getting a lot worse scores? You're not running in very high CVSiN or something else is not set to max / ON.

A friend of mine runs the game on a Q6600 @ 3GHz and a GTX570, the newest newest betas. With the high rez texture pack, at 1920x1200. Setting all to max, no AA and:

-> High = smooth and great performance (40-50FPS)
-> Very High = 20-25 FPS

Looks like you're running the game in High, CVSiN.

 

[RobertR1]

What does that have to do with him waiting for a fix? Pissed or not?

Wouldn't you wait until a fix was available for a much better gaming experience? Or would you plod through it when something isn't right?

People buy games on day 1 so they can play them on day 1. Same reason people do preloads and midnight pickups. They want to play the game as early as possible after all the anticipation.

With his setup, it's reasonable to assume that he could buy any game day 1 and play it cranked up with great performance.

 

[Madcatatlas]

I regulerly buy games and just put them on the shelve, waiting for them and drivers to mature. Like wine almost, if you know what i mean.


On a completely related note. What engine IS Dragon Age 2 using, and are there more games coming this year that use the same engine?

The issue might be bigger than DA2

 

[Vette73]

,.

 

[alcoholbob]

CVSiN is getting 50FPS, yet everyone else, even people having GTX580 SLi (or GTX480x3!) are getting a lot worse scores? You're not running in very high CVSiN or something else is not set to max / ON.

A friend of mine runs the game on a Q6600 @ 3GHz and a GTX570, the newest newest betas. With the high rez texture pack, at 1920x1200. Setting all to max, no AA and:

-> High = smooth and great performance (40-50FPS)
-> Very High = 20-25 FPS

Looks like you're running the game in High, CVSiN.

There was a guy on the Bioshock forums with a 3 year old laptop reporting maxing out Bioshock when it came out, at a time where every review on the web had top end systems barely capable of 40fps.

So what can you deduce? People online will say anything. I won't go so far to say people are trolling or lying about maxing the game...because I'm not the confrontational type. If you are an intelligent person you can come to your own conclusions about people's claiming apples fall upwards.

 

[SHAQ]

It runs bad on very high even with all the DX11 settings turned off and AA so there is a driver bug somewhere.

 

[RussianSensation]

5770 (40 fps) > 460 1GB (30 fps) fps wise atm for Dragon Age II in directX 10.

It just needs some driver fixing from nvidia, and things will probably get better.

But if you see that 3870 > 8800GT, HD4870 is almost doubling GTX260 216, then it's starting to look like this game simply favours AMD's architectural design. How can GF8, GT200 and Fermi all perform worse than AMD's comparable generations? This actually sends a signal it is not a driver issue.

In fact, this reminds me exactly of Civilization 5 where GT200 architecute was simply outclassed based on architectural bottlenecks for that game's specific code: http://www.techspot.com/review/320-civilization-v-performance/page6.html

However, Civ5 runs beautifully on Fermi architecture, yet an HD4890 is almost 2x faster than a GTX260 216.

The fact that its a problem in dx10 mode as well for NV cards suggest this is a bigger problem and may have to do with the game engine itself.

Bingo. A DX11 driver issue/bug related to Fermi should not have any effect on an HD4870 almost doubling a GTX260 in DX10. Oh well, NV smokes AMD in Lost Planet 2. You win some, you lose some. It just happens that DA2 is a much better game.

It runs bad on very high even with all the DX11 settings turned off and AA so there is a driver bug somewhere.

It's possible, but I doubt GTX580 can overcome a 46% deficit compared to the HD6970 even with a driver fix.

 

[Skurge]

But if you see that 3870 > 8800GT, HD4870 is almost doubling GTX260 216, then it's starting to look like this game simply favours AMD's architectural design. How can GF8, GT200 and Fermi all perform worse than AMD's comparable generations? This actually sends a signal it is not a driver issue.

In fact, this reminds me exactly of Civilization 5 where GT200 architecute was simply outclassed based on architectural bottlenecks for that game's specific code: http://www.techspot.com/review/320-civilization-v-performance/page6.html

However, Civ5 runs beautifully on Fermi architecture, yet an HD4890 is almost 2x faster than a GTX260 216.



Bingo. A DX11 driver issue/bug related to Fermi should not have any effect on an HD4870 almost doubling a GTX260 in DX10. Oh well, NV smokes AMD in Lost Planet 2. You win some, you lose some. It just happens that DA2 is a much better game.



See above.

You know, this has been a pattern I've been noticing. AMD's older DX10 GPUs (RV670, RV770) Do well in modern DX10 games whereas Nv's older DX10 GPUs (G8x, GT200) Do poorly. Could it just be AMD's architecture was more forward looking than nVidia's. Just like R580 vs G70. Or Do AMD support older cards better in their drivers?

Also when you turn off DX11 in Lost planet, things go back to parity with AMD vs Nv, not so for dragon age. Could Just be a game AAMD will always be faster in.

AMD also helped with SHOGUN 2 DX11 implementation. Could we be seeing the same thing happen again?

Is AMD Purposely crippling performance on Nv cards? Or is optimizing the game code for the architecture causing nV cards to perform extremely poorly?

 

[RussianSensation]

You know, this has been a pattern I've been noticing. AMD's older DX10 GPUs (RV670, RV770) Do well in modern DX10 games whereas Nv's older DX10 GPUs (G8x, GT200) Do poorly.

There are cases where NV's cards just run faster too though. Like BlackOps where a GTX275 is a solid 57% faster than a 4890. Generally speaking though, I agree with you that AMD's cards tend to "age" better with time. This has been true for several generations actually (outside of 2900XT and 3870).

For insance, X1900XT 256mb gives a 7900GTX 512mb a beating and a half in Call of Duty Modern Warfare 1. Back in the days a 7900GTX was not that much slower.

Also when you turn off DX11 in Lost planet, things go back to parity with AMD vs Nv, not so for dragon age.

In LP2 though, DX11 brings in extreme tessellation. This is probably why NV's cards outperform AMD's under that setting. Ironically, it hardly improves the visuals.

AMD also helped with SHOGUN 2 DX11 implementation. Could we be seeing the same thing happen again? Is AMD Purposely crippling performance on Nv cards? Or is optimizing the game code for the architecture causing nV cards to perform extremely poorly?

We have seen what code optimization for a specific hardware can do on consoles. It's very possible that AMD worked very closely with Bioware to optimize the code specifically for their highly parallel architecture. It's a fair game.

NV pays $ to developers to crank Tessellation to "64 billion depth". Wait until Crysis 2 DX11 patch. :awe: AMD's cards may run faster with some of these effects: SSAO, bloom, depth of field. (see HD6990s slide to see the specific features). It's just exploiting your architecture's strengths. We won't know for sure until NV has had a fair chance to improve their drivers. I was always under the impression that NV had a superior DX11 architecture but DA2 makes me question that opinion.

I haven't seen Shogun 2 DX11 benches. Do you have a link?

 

[Skurge]

There are cases where NV's cards just run faster too though. Like BlackOps where a GTX275 is a solid 57% faster than a 4890. Generally speaking though, I agree with you that AMD's cards tend to "age" better with time. This has been true for several generations actually (outside of 2900XT and 3870).

For insance, X1900XT 256mb gives a 7900GTX 512mb a beating and a half in Call of Duty Modern Warfare 1. Back in the days a 7900GTX was not that much slower.



In LP2 though, DX11 brings in extreme tessellation. This is probably why NV's cards outperform AMD's under that setting. Ironically, it hardly improves the visuals.



We have seen what code optimization for a specific hardware can do on consoles. It's very possible that AMD worked very closely with Bioware to optimize the code specifically for their highly parallel architecture. It's a fair game.

NV pays $ to developers to crank Tessellation to "64 billion depth". Wait until Crysis 2 DX11 patch. :awe: AMD's cards may run faster with some of these effects: SSAO, bloom, depth of field. (see HD6990s slide to see the specific features). It's just exploiting your architecture's strengths. We won't know for sure until NV has had a fair chance to improve their drivers. I was always under the impression that NV had a superior DX11 architecture but DA2 makes me question that opinion.

I haven't seen Shogun 2 DX11 benches. Do you have a link?

http://www.hardwareheaven.com/revie...-graphics-card-review-total-war-shogun-2.html

I only have this but as you can see, SLI/crossfire doesn't work in that game. There are also some reports that, the game is only getting a DX11 patch in a few weeks. So I don't know if those benches are relevant.

 

[RussianSensation]

http://www.hardwareheaven.com/revie...-graphics-card-review-total-war-shogun-2.html

I only have this but as you can see, SLI/crossfire doesn't work in that game. There are also some reports that, the game is only getting a DX11 patch in a few weeks. So I don't know if those benches are relevant.

Ok but that looks totally different. A single GTX580 is easily beating a single 5870/6970 in that bench (considering CF isn't working). So that's clearly a driver issue that will be fixed. In DA2, every single NV generation loses to a comparable AMD generation. Probably SSAO+DOF is killing NV's cards. I find it interesting no one has tested all of these intensive graphical settings one by one to figure out which one(s) specifically hammer their NV cards. :hmm:

 

[Skurge]

Ok but that looks totally different. A single GTX580 is easily beating a single 5870/6970 in that bench (considering CF isn't working). So that's clearly a driver issue that will be fixed. In DA2, every single NV generation loses to a comparable AMD generation. Probably SSAO+DOF is killing NV's cards. I find it interesting no one has tested all of these intensive graphical settings one by one to figure out which one(s) specifically hammer their NV cards. :hmm:

Apparently they were using the demo of the game, so I don't think that benchmark should be taken seriously.

 

[alcoholbob]

You know, this has been a pattern I've been noticing. AMD's older DX10 GPUs (RV670, RV770) Do well in modern DX10 games whereas Nv's older DX10 GPUs (G8x, GT200) Do poorly. Could it just be AMD's architecture was more forward looking than nVidia's. Just like R580 vs G70. Or Do AMD support older cards better in their drivers?

Also when you turn off DX11 in Lost planet, things go back to parity with AMD vs Nv, not so for dragon age. Could Just be a game AAMD will always be faster in.

AMD also helped with SHOGUN 2 DX11 implementation. Could we be seeing the same thing happen again?

Is AMD Purposely crippling performance on Nv cards? Or is optimizing the game code for the architecture causing nV cards to perform extremely poorly?

I can guarantee you this is not the case. AMD hardware is running relatively better than Nvidia hardware but with all the details on max it's still utter ass. You get better performance maxing out Crysis...

 

[solarissf]

ok guys... so I'm trying to follow along... I have a 560ti, and cannot play on very high. after all this has there been a solution? or still in the works?
I did download the drivers from nvideo which were released mar1 but to no avail

 

[Qbah]

I can guarantee you this is not the case. AMD hardware is running relatively better than Nvidia hardware but with all the details on max it's still utter ass. You get better performance maxing out Crysis...

Heh, that is actually very true. I'm getting an average of 30FPS on Crysis all enthusiast and AAx4. DA2 gives me around ~25 FPS on average all maxed and AAx4. So yeah, it's not like AMD owners get a buttery smooth experience either.