- Joined
- Mar 8, 2007
- Messages
- 4,098
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
HD 2900XT & Driver 8.38
- Thread starter Eastcoasthandle
- Start date
Meaningless, without a valid, reputable and creditable source.
Here you go:
http://enthusiast.hardocp.com/article.html?art=MTExLDEsLGhlbnRodXNpYXN0
http://www.techreport.com/etc/2001q4/radeon-q3/index.x?pg=1
You and a couple others in this thread, are no doubt, the group that calls it a "driver bug", Yeah, sure it was a driver bug. No optimization at all.
Just read those articles and understand this. It doesn't matter which company does it (NVIDIA or ATI), it's a cheat and they got flamed for it and they should be. If you're supporting their cheat, they won't feel compelled to actually stop doing it. ATI did, NVIDIA did it and they both had to remove the cheat or it means bad publicity for them and as a money making company, they don't want that.
Eastcoasthandle
[H]ard|Gawd
- Joined
- Jan 27, 2006
- Messages
- 1,041
One would have to changing quake 3 to Quack 3, is that the best you can do?Here you go:
http://enthusiast.hardocp.com/article.html?art=MTExLDEsLGhlbnRodXNpYXN0
http://www.techreport.com/etc/2001q4/radeon-q3/index.x?pg=1
You and a couple others in this thread, are no doubt, the group that calls it a "driver bug", Yeah, sure it was a driver bug. No optimization at all.
Just read those articles and understand this. It doesn't matter which company does it (NVIDIA or ATI), it's a cheat and they got flamed for it and they should be. If you're supporting their cheat, they won't feel compelled to actually stop doing it. ATI did, NVIDIA did it and they both had to remove the cheat or it means bad publicity for them and as a money making company, they don't want that.
Oh wait, you simply made a baseless comparison with the HD 2900XT and the 8500 on 1 driver issue. It's obvious you are posting FUD. LOL
One would have to changing quake 3 to Quack 3, is that the best you can do?
Oh wait, you simply made a baseless comparison with the HD 2900XT and the 8500 on 1 driver issue. It's obvious you are posting FUD. LOL
You're obviously a troll, so I'll just say this before I ignore you.
Both articles mention that if you run the game as quake3.exe, there are obvious optimizations being done and textures are all washed out (detail is much, much lower). Now if you change the name of the game to something else other than quake3.exe (and all its instances in the executable), you'll notice that everything looks ok, though with a large performance hit.
That is a more than obvious optimization to quake3, which means in the driver there's a check to verify if the game is quake3 and if it is, it will reduce the amount of detail rendered, therefore increasing performance. And no, that's no bug.
Eastcoasthandle
[H]ard|Gawd
- Joined
- Jan 27, 2006
- Messages
- 1,041
Here you go:
http://enthusiast.hardocp.com/article.html?art=MTExLDEsLGhlbnRodXNpYXN0
http://www.techreport.com/etc/2001q4/radeon-q3/index.x?pg=1
You and a couple others in this thread, are no doubt, the group that calls it a "driver bug", Yeah, sure it was a driver bug. No optimization at all.
Just read those articles and understand this. It doesn't matter which company does it (NVIDIA or ATI), it's a cheat and they got flamed for it and they should be. If you're supporting their cheat, they won't feel compelled to actually stop doing it. ATI did, NVIDIA did it and they both had to remove the cheat or it means bad publicity for them and as a money making company, they don't want that.
What does Quake 3 have to do with HD 2900XT? <nothing>.
You are posting FUD here because there is no association with the HD 2900XT and quake 3. Although you are trying to imply some association based on a driver problem nearly 6 years ago, lol. The only one trolling is you!
how about the color mip map detection of the x800's? Or did you forget that one Eastcoasthandle?
Now I don't know about the final release drivers coming out this month but some of the rc drivers and beta drivers as of late do have some questionable things going on, bugs possibly, since I'm not testing them myself, its hard to just say something, but some funny things were going on.
That said, this is a pointless arguement just wait till the final drivers are out and then do some investigation.
Now I don't know about the final release drivers coming out this month but some of the rc drivers and beta drivers as of late do have some questionable things going on, bugs possibly, since I'm not testing them myself, its hard to just say something, but some funny things were going on.
That said, this is a pointless arguement just wait till the final drivers are out and then do some investigation.
Eastcoasthandle
[H]ard|Gawd
- Joined
- Jan 27, 2006
- Messages
- 1,041
If it's pointless why is it important to you to post about it? Look, what was recently released is the HD 2900 XT. Right now there are some driver issues that needs to be addressed. It may take them more then just 1 driver release to fix the problems. Pardon me if I don't agree with some conspiracy theory or pity party.
Oh, BTW what about the release of the G80 and vista problems? Or did you forget that razor1?
Ok, you want to talk about cheats? Or do you want to talk about driver bugs? Or do you want me to say there is no way the 2900xt will get much more performance with the next drivers? That last question, is actually a statement, because there are questionable things going on. Its possible its a bug, but I really don't think so, when certain AA sampling algos are swapped for other. That is not a bug, no way in hell can that be, because it seems to be on a per game basis. Specific games that seem to this, Oblivion and TDU, where it would have the most impact. Now I don't know if this is happening with the final release it is happening with some of the earlier drivers mentioned in this thread a few posts ago though.
Oh btw, what did ATi have out at the time when nV released the g80? Oh I forgot nothing..... so does it really matter much now? And the g80 had vista drivers when Vista was launched, they might not have been best of drivers but they were there and a month after they got the stability pretty much as needed. The people that dled the RTM version or got the RTM version legally, tough luck, Vista wasn't released to the public at that time, and there is bound to be issues. I guess you don't remember all the issues with Win Xp when it was released? Its going to happen no matter what, its a new OS, it takes time to get drivers ready, but I don't really care about the bugs, its when you have selected options changing that peaks my interest.
Lets see those better ATi vista drivers now they said they had then 4 months ago, well crossfire wasn't working that great, and you are now saying they are still buggy for single cards too?
Eastcoasthandle
[H]ard|Gawd
- Joined
- Jan 27, 2006
- Messages
- 1,041
Your post gives the appearance that you are guessing.
-You want to talk about cheats
-Next, it's about bugs
-Next, you are not sure
-Next, you think it's AA sampling Algos
-Now it's not a bug issue
-Next, it's on a per game bases
Oh, no need to be touchy about the vista issue, it did happen regardless if the HD 2900XT was released during that time or not.
.Your post gives the appearance that you are guessing.
-You want to talk about cheats
-Next, it's about bugs
-Next, you are not sure
-Next, you think it's AA sampling Algos
-Now it's not a bug issue
-Next, it's on a per game bases
Oh, no need to be touchy about the vista issue, it did happen regardless if the HD 2900XT was released during that time or not.
might want to edit your post, since I editted mine
You have no idea whats going on with these drivers anyways I doubt you ever will because you might want to dl the 3872 drivers and test it out for yourself they have been out there, since you are so supportive of all the problems being bugs, you will see some changes that don't seem to be bugs. Just search the web for the drivers you will find them they have been leaked. (this is about per game basis, SSAAA sampling, these aren't bugs, wow that covered everything right) and yeah its not a big issue because they aren't released versions of the driver, there is an issue some review sites are using the drivers yes, but wait and see till the final drivers are released to make the judgement call,
ATi has played review sites before thats why I stated the mip map detection where they specifically told reviewers to do the color map map tests but didn't tell them they were detecting when that program was being used and end results gave an output of something better then it really was. Coicidence isn't it? I guess you didn't understand where I was coming from with mip map detection?
Eastcoasthandle
[H]ard|Gawd
- Joined
- Jan 27, 2006
- Messages
- 1,041
might want to edit your post, since I editted mine
You have no idea whats going on with these drivers anyways I doubt you ever will because you might want to dl the 3872 drivers and test it out for yourself they have been out there, since you are so supportive of all the problems being bugs, you will see some changes that don't seem to be bugs. Just search the web for the drivers you will find them they have been leaked. (this is about per game basis, SSAAA sampling, these aren't bugs, wow that covered everything right) and yeah its not a big issue because they aren't released versions of the driver, there is an issue some review sites are using the drivers yes, but wait and see till the final drivers are released to make the judgement call,
ATi has played review sites before thats why I stated the mip map detection where they specifically told reviewers to do the color map map tests but didn't tell them they were detecting when that program was being used and end results gave an output of something better then it really was. Coicidence isn't it? I guess you didn't understand where I was coming from with mip map detection?
Looks like the G80 has problems of it's own
I guess it doesn't pay to slight one video card manufacture over another.
Looks like the G80 has problems of it's own
I guess it doesn't pay to slight one video card manufacture over another.
My god more then half of those 12 problems aren't even problems with the g80 or have been fixed, read the entire post, did you?
1,2,7,9,10
These problems are there for the x2900xt or are problems out side of the g80.
3,12
These problems have been fixed, I've been overclocking my g80 since the 158 drivers with ntune. In XP you could overclock right off the bat, I never had any issues in XP even with the early drivers.
THen you have X-Fi, I have this and tell ya what, I've bitched out Creative a few times where are their good Vista drivers and Software? At least the damn thing works somewhat, but still needs alot more work, no problem with XP at all though.
thats 8/12 problems that are cured, or not problems with the g80, so what do you have left I don't know much about since I haven't experienced any issues of the sort.
As I said in another thread, I think there are individuals here who post solely to entrap other users. I don't really understand why, but I'm fairly sure this is the situation, and I believe Eastcoasthandle is one of those individuals.
My advice is to just ignore it. It's not as if anyone will be buying a card based on the incoherent musings of some guy on some forum anyway.
My advice is to just ignore it. It's not as if anyone will be buying a card based on the incoherent musings of some guy on some forum anyway.
Anarchist4000
[H]ard|Gawd
- Joined
- Jun 10, 2001
- Messages
- 1,659
I think that really depends on the guy and the forums involved. But anything posted on this forum should probably be taken with a couple tons of salt.
Its pretty funny NV fans in an ATi subforum, telling others to ignore trolls. When its NV fans doing most of the trolling. Heh.
Sorry, it was a bug. Which was explained in depth afterwards. Backed up by facts, when a driver released less than a week later, fixed the problem, and raise frames. Believe what you want, the facts are the facts. Its not like [H] has never been wrong. I guess you know nothing about the P4 debacle.
Here you go:
http://enthusiast.hardocp.com/article.html?art=MTExLDEsLGhlbnRodXNpYXN0
http://www.techreport.com/etc/2001q4/radeon-q3/index.x?pg=1
You and a couple others in this thread, are no doubt, the group that calls it a "driver bug", Yeah, sure it was a driver bug. No optimization at all.
Just read those articles and understand this. It doesn't matter which company does it (NVIDIA or ATI), it's a cheat and they got flamed for it and they should be. If you're supporting their cheat, they won't feel compelled to actually stop doing it. ATI did, NVIDIA did it and they both had to remove the cheat or it means bad publicity for them and as a money making company, they don't want that.
Sorry, it was a bug. Which was explained in depth afterwards. Backed up by facts, when a driver released less than a week later, fixed the problem, and raise frames. Believe what you want, the facts are the facts. Its not like [H] has never been wrong. I guess you know nothing about the P4 debacle.
I think it's reasonable to assume that it was an incident on ATi purposefully degrading image quality via application detection to improve performance, and the subsequent drivers reversed these optimizations (while improving performance as you said). ATi's explanation was that it was a driver bug, but as far as I'm aware, there are no "facts" that would indicate that it wasn't cheating.
Of course NVIDIA's been guilty of the same thing who-knows-how-many times (and in some of those cases, they offered the same explanations), and of course the 8500 has nothing to do with the HD 2900. I quite frankly don't know why the issue got brought up, and why an association was made with the 2900.
It's still very early, and all of these issues can be (and rightly should be) resolved. I don't think It's a question of AMD "cheating" so much as it is a question of hardware reviewers either not paying any attention or deliberately attempting to be misleading.
And, for reference, I never said that it was just ATi fans doing the trolling here, fallguy.
Of course NVIDIA's been guilty of the same thing who-knows-how-many times (and in some of those cases, they offered the same explanations), and of course the 8500 has nothing to do with the HD 2900. I quite frankly don't know why the issue got brought up, and why an association was made with the 2900.
It's still very early, and all of these issues can be (and rightly should be) resolved. I don't think It's a question of AMD "cheating" so much as it is a question of hardware reviewers either not paying any attention or deliberately attempting to be misleading.
And, for reference, I never said that it was just ATi fans doing the trolling here, fallguy.
alg7_munif
Supreme [H]ardness
- Joined
- Oct 9, 2006
- Messages
- 5,862
I think that the driver can still be improved for performance since the new card is an unified shader card but it is used to play games that are developed for cards with seperated pixel pipeline and vertex shader but don't expect that it would be too high.
So let me get this straight, the 8500 release is well known for its horrible drivers. And then there happens to be a bug going from the R100 to the R200 in the drivers that screws up with IQ on a game. Then 5 days (Im pretty sure, not 100%
later a driver comes out and fixes the issue, and increases fps, and its still cheating? Sorry, proof is in the pudding. Increased frames, while fixing the issue. If they degraded IQ to get more frames, then how did they get more frames while fixing the issue so fast. As I said, the 8500 launch was horrible driver wise. They had many, many issues. I dont see why its so hard to believe that this problem was a bug, when there was so many so begin with. With several driver releases, they increased performance greatly. Catching, and passing NV. In both IQ and FPS.The same thing? No. NV was flat out caught in 3Dmark. No bug about it.
Thats great. As long as we're on the same lines that its mostly NV fans, doing nothing but trolling in an ATi subforum. Too many are getting upset that the 2900XT will get faster with newer drivers. Boo-hoo. That should be good news for everyone.
Who's talking about 3DMark?
I can't imagine anyone getting upset over that so long as there is no IQ degradation. I think it's important that we keep an eye out for it, but I don't think AMD intends to pull any cheap tricks.
- Joined
- Mar 8, 2007
- Messages
- 4,098
Today I installed the leaked 8.38 RC3 drivers and I did not notice any appreciable increase in performance. All I know is that Oblivion is not impressive with 4xaa, it lags, Battlefield2 is almost unplayable with 8xaa enabled, and Half Life 2 just doesn't look right.
I hope the official drivers are better.
I hope the official drivers are better.
No one made any association with 8500 and 2900 XT. It was Eastcoasthandle that doubted the 8500 driver optimization and asked for proof. I gave it to him.
fallguy said:
I'm actually amazed at how you <insert company name here> fans, believe in everything that company tells you. Yeah, it was a bug. The driver would just pick up the "quake3" keyword and due to that bug, it would decrease the amount of detail that the card needs to render and therefore increasing fps.
Does this sound about right ? I'm a software developer and that is NOT something you do by mistake
And again, to emphasize my point earlier. It doesn't matter which company does it. If you don't flame them for doing it, they won't feel compelled to stop doing it at all.
Please stop the flag waving. Both companies are the same. They don't care if it's right or wrong as long as people don't notice it and their cards sell.
fallguy said:
First of all, any user is entitled to visit any sub-forum as he/she sees fit, as long as he/she maintains an adult discussion,
Second, it's actually you who seems quite upset when someone even raises the question that the new drivers are not working properly (i..e not rendering HDR as someone mentioned in this thread and originated the IQ discussion).
Thanks for the input. With all this discussion, which I was also part of, I admit I disregarded your posts, which are probably the most valuable ones, since you have a HD 2900 XT to test on. I'm sorry for that
Eastcoasthandle
[H]ard|Gawd
- Joined
- Jan 27, 2006
- Messages
- 1,041
Lies, I tell you...liesSilus said:
You are the one that added to the conversation making you an accomplice to this conspiracy theory (HD 2900XT is like the 8500, etc). Below you will find who made the association between the 8500 and 2900XT. Read below, I suggest in the future you actually research before you falsely accuse people. It doesn't help your credibility here.APOLLO said:
InorganicMatter said:
Lies, I tell you...lies
I don't even know why I'm replying to you, but I guess I'll play your game this time.
Eastcoasthandle said:
I gave you the source. Sources in fact, which in your troll like form you refuted, even though there's not much room to doubt what's written in those articles.
Eastcoasthandle
[H]ard|Gawd
- Joined
- Jan 27, 2006
- Messages
- 1,041
Lies, I tell you...liesSilus said:
You are the one that added to the conversation making you an accomplice to this conspiracy theory (HD 2900XT is like the 8500, etc). Below you will find who made the association between the 8500 and 2900XT. Read below, I suggest in the future you actually research before you falsely accuse people. It doesn't help your credibility here.APOLLO said:
InorganicMatter said:
Silus, your post (above this post) has nothing to do with the facts I presented, please stay on topic. The issue here is that you said that no one made any association with the 8500 and 2900XT. The quotes above prove this is not true. Also, you did make yourself an accomplice when you added to the association that the 8500 was like the 2900XT when you provided those links (conspiracy theory).
2nd, calling me a troll because I don't agree with your point of view, ask for proof, etc only discredits your use of the word because it's not used correctly. Please learn how to apply the term before using it.Silus said:
ShiningHolden
Limp Gawd
- Joined
- Oct 21, 2006
- Messages
- 315
I was going to 8800GTX SLI, but I hate the nForce 680i SLI chipset enough to consider CrossFire X2900XT.
Did 8.38 drivers make the HD 2900XT's better than in games?
Did 8.38 drivers make the HD 2900XT's better than in games?
So far it looks like they fixed some stuff but will take another release or two to see what we have.
ShiningHolden
Limp Gawd
- Joined
- Oct 21, 2006
- Messages
- 315
Well, if the Intel P35 chipset can be modded to run SLi like the ASUS P5B was with patches and drivers, and the performance doesn't suck the wazoo, then I am going to be doing some modding.
I just don't want the nForce 600 series.Are the 8.38 's OFFICIALLY released?
What do we have to wait for, 8.39?
Well, if the Intel P35 chipset can be modded to run SLi like the ASUS P5B was with patches and drivers, and the performance doesn't suck the wazoo, then I am going to be doing some modding.
Are the 8.38 's OFFICIALLY released?
What do we have to wait for, 8.39?
AT's write-up on P35 X-Fire performance had it trailing the 975X in most cases, sometimes by sustantial margins. SLi will almost certainly behave similarly, assuming the chipset can be tricked into running it.
As for drivers, it looks like AMD have just as many issues with their R600 drivers as nV had with G80 at launch. So I'd guess there will be several Catalyst releases before things are relatively peachy on the driver front. High shader loads and AA are always going to be a problem due the lack of a dedicated multisampling unit. I read at a couple of sites that it was borked (this was before release), and to me that's more logical than AMD deciding it wasn't needed. The X1900's really handle AA very well, and you'd think they'd be keen to keep hardware resolve considering review sites and users nearly always want 4x enabled. Stating a "forward-looking design" is better PR than admitting "hey, we broke it. Sorry"
Heck, I like to think it's borked - at least then there is a chance of a refresh with a working unit
Considering performance with AA disabled, the R600 would clearly be quite an eye-candy animal were it not for this problem (or design flaw, whichever is true).
According to Gary Keys of Anandtech, the release date for Cat 7.5 is May 31st.
I dont believe anything a company tells me. I guess you missed the part where a driver 5 days later, fixed the issue, and increased frames. Neverminding the fact that NVs skybox looked like crap compared to ATis. Obviously they didnt need to cheat. Once again, people have no problem admitting that the 8500 had horrible, horrible drivers, with many bugs, yet cant seem to think this one was one as well. "I'm actually amazed at how you <insert company name here> fans, believe in everything negative for the other brand."
I didnt wave any flag. I simply said comparing a bug for ATi, and NV's blatant attempt at getting higher scores thru a false way in 3dMark, are not the same.
Sure they are. I just pointed out that its funny all these NV fans in an ATi subforum getting upset that the 2900XT is getting better with newer drivers. Its a fact that its going to happen, you need to get over it. All this post was, was an update with new drivers, and how that its performing better. Yet NV fans didnt like it, and posted negatives about it. Boo-hoo.
Your assessment of the thread is about as far off as you can possibly get.
Firstly, you're spinning reality here. I didn't see anyone commenting that they were upset that the 8.38s are improving performance. That's an absurd interpretation. What I did see, however, was some questioning of the benchmarks. Caff brought up some SM2.0/SM3.0 discrepancies which he thought would also negatively influence the rest of the benchmarks. Given the test machine, and the drivers, the discrepancy is well worth questioning. So-called "NV fans" were upset about the benchmarks, not about AMD or 8.38.
Then, we got a link to the PCPOP benchmarks, and again there were some issues. The playing field wasn't even here -- there are obvious visual discrepancies, which Eastcoasthandle said did not exist (in fact being so bold as to say that they're "fine", and told j to stop being picky). The 2900 XT isn't rendering scenes correctly, omitting HDR completely it seems in some cases, which would severely inflate any 2900 numbers. PCPOP's other numbers might be totally fine to look at for comparison, and would be valuable to get a better idea of 8.38 performance, but not alerting people to the very large differences in rendering in said benchmarks is neglectful, don't you think? If it caught your eye, would you say anything? Would you let it pass?
I'm sure the 8.38s improve performance, but it makes sense to invalidate certain aspects of reviews because of questionable numbers or IQ. You'd do the same were it a NVIDIA card under discussion. That's kind of why we're here, you know; to look at data much more closely than your average Anandtech kid (or whatever). So, people look at that data, comment on it, and your assertion is that they're unhappy that AMD is making headway? There are trollish posts throughout, no doubt, but there aren't that many. The bulk of the thread consists of reasonable points and mature and semi-mature discussion.
Exactly phide, I could care less which card is actually better. But right now, the 8800gts 640mb is faster in most games (that are rendered correctly on both cards) and is cheaper. I would recommend the 2900 if it performed better than the 8800gts, but it doesn't, and its more expensive.
Will drivers make it better? Maybe. Hopefully for the price point, but we don't know 100% that performance will make up the huge gap in price/performance for the two cards, and people need to know that before they spend $400+ on it.
I know if I just spend $400 on my card, and it looked like crap and ran as slow as my last gen card I'd be PISSED. People need to let ATI/AMD know that they won't put up with this kind of quality from such expensive pieces of hardware. Just saying "oh it will get better" takes away from the fact that the company screwed up, and letting them get away with that will only lead to them doing it again.
And ATI/AMD has had basically the SAME amount of time as Nvidia has for this card and drivers, so there really is no excuse.
Will drivers make it better? Maybe. Hopefully for the price point, but we don't know 100% that performance will make up the huge gap in price/performance for the two cards, and people need to know that before they spend $400+ on it.
I know if I just spend $400 on my card, and it looked like crap and ran as slow as my last gen card I'd be PISSED. People need to let ATI/AMD know that they won't put up with this kind of quality from such expensive pieces of hardware. Just saying "oh it will get better" takes away from the fact that the company screwed up, and letting them get away with that will only lead to them doing it again.
And ATI/AMD has had basically the SAME amount of time as Nvidia has for this card and drivers, so there really is no excuse.
- Joined
- Mar 8, 2007
- Messages
- 4,098
Thanks for the input. With all this discussion, which I was also part of, I admit I disregarded your posts, which are probably the most valuable ones, since you have a HD 2900 XT to test on. I'm sorry for that
I have this card guys and I've done a lot of research on it. It was hard for me to get my mind around this for some reason but the fact of the matter is that ATi build a next gen card without any advances in antialiasing performance. When you get into any kind of situation where there is a lot of foliage or trees on screen the card slows to a crawl with anything greater than 4x enabled, and in Oblivion anything greater than 2x enabled.
Frame rates are not impressive. In Battlefield 2, 4xaa 17xaf, frame rates dip into the 40s in some areas. This is way too low.
I think the issue is the back end--the lack of ROPs. ATi has focused on shader based aa, or programmable aa, which is basically a feature that game developers have not yet implemented. The lack of ROPs (which handle the aa) is either a flaw in the design or a conscious design choice based on assumptions ATi is making about what kind of architecture will be most efficient with future games.
The one thing that I am sure of, without a shadow of a doubt, is that this card was not designed to run games on the market as of its release--Oblivion, ect. The way this card was designed, it's as if ATi considers those games to be "legacy games". But we are several months from the release of any killer DX10 app, and we can only judge the card based on what we can play today. I've been playing BF2, Oblivion, Supreme Commander, Unreal Tournament 2004, Counterstrike, and Half-Life 2 and I'm just not impressed. Either the frame rates are not good or things just don't look good--there is shimmering, things don't render right, ect.
In Oblivion, I posted a thread here showing screen shots of the same scene with each of the 2900's 9 anti aliasing modes engaged. For reasons I can't explain, canopy shadows disappeared with anything higher than the first mode--they just totally vanished.
There is a lot of fishy stuff going I do not recommend this card. I think the Nivida cards are the way to go, or wait for a better ATi card, but I just find it hard to believe that drivers are going to save this card from its poor aa performance given that we know it lacks ROPs.
Note that I am using driver 8.38 RC7.
See my other threads:
http://www.hardforum.com/showthread.php?t=1195344
http://www.hardforum.com/showthread.php?t=1195556