ATI hand in hand with NVidia?

[SurfMonkey]

Looks like Unwinders pulling apart the Catalyst driver set now. And he's found some pretty interesting things.

Like a huge drop in framerates in that nature scene of 3dmark2001 if the cats are stopped from detecting the application.

Hmmm, can we say clipping? Shader replacement, big difference in image quality before and after.

I guess we wait for the articles to appear on the web before the fighting starts.

But if it's true, then ATI are just as bad as NVidia. Both are willing to distort the truth in order to sell cards.

Source (http://www.beyond3d.com/forum/viewtopic.php?t=6480)

[digitalwanderer]

Originally posted by SurfMonkey
Looks like Unwinders pulling apart the Catalyst driver set now. And he's found some pretty interesting things.

Like a huge drop in framerates in that nature scene of 3dmark2001 if the cats are stopped from detecting the application.

Hmmm, can we say clipping? Shader replacement, big difference in image quality before and after.

I guess we wait for the articles to appear on the web before the fighting starts.

But if it's true, then ATI are just as bad as NVidia. Both are willing to distort the truth in order to sell cards.

Source (http://www.beyond3d.com/forum/viewtopic.php?t=6480)
I'm waiting to see how ATi reacts to this news and what they say about it before casting judgement, and I'm STILL waiting for nVidia to comment on ANY of this!

ATi has handled their cheats with a lot more class and honesty than nVidia has so far, I expect that will be the case in this one too.

[deejaya]

Well I thought it was bad with 3dMark03, both companies, and ATi come up with some PR (whether true or not) to counter it. While better than nVidia's response, it doesn't mean they didn't do it. Now they have both been caught in 2k1. Even if both companies came out with PR about it being legitimate, it doesn't change anything.

They both cheated in popular benchmarks to decieve the average Joe, as everyone keeps saying.

[saturnotaku]

Originally posted by digitalwanderer
ATi has handled their cheats with a lot more class and honesty than nVidia has so far, I expect that will be the case in this one too.

But if this was going on at the same time as when ATI admitted they were "cheating" in 3dmark03, just how honest have they been? :wtf:

[Skynet]

We can all argue endlessly (and have) over Nvidia cheating, ATI cheating etc. The bottom line is NOTHING is going to change the fact that Nvidia's FX cards have terrible shader performance. Even the mighty 5900 falls face down in PS performance. For some reason this is totally lost in the debate. I have not had a chance to test a 5900 but did put a 5600 in a system a few days back and was very dissapointed in it. With eye candy enabled it was pretty slow. The 9500 wipes the floor with this card, although the 9600 dissapointed me as well.

IMHO the 9700, 9700pro and of course the 9800 are far and away the best overall graphics card on Earth.

[GlowStick]

*sarcasm*
Hey guys, a 66->42 FPS decrease when app detection is blocked IS NOT CHEATING. They are just optmizing, it is OK for ATi to do it, its PREFECTLY OK.

ATi is handling this so much better, after they were caught cheating, they said yeah we were cheating only in 3dmark03, now that they were caught again in 3dmark01 they will probly admit it, that is so much better than cheating and not telling anyone about it untill someone catches you, like Nvidia did.
*/sarcasm*

omg, ati infact is handling this exactly like Nvidia, they wont fess up to it untill confronted, then claim its just an optimzation.

[StealthHawk]

As stated http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=13217&pagenumber=7

AT least 3 pixel shaders are detected (1@1.1 and 2@2.0), texture detections (thats how GT4 in 3dmark2001 is detected) ....
Full texture patterns& shader code are in the driver, but although easy to see what they do, its very hard to make FULL anti-detect patch.....
.... what ATi make with these 2 p.sh. 2.0 is hardly "shifting" ......

This "optimization" is a lot more "shady" than what ATI did with 3dmark03 ;)

[DSC]

Any threads started at Rage3d on this or comments from the fanATIcs at Rage3d about this? ;) :D

[Star_Hunter]

Yes there are but plz keep in mind untill we get more detail in terms of how he did it we dont know if it truely a cheat or not there are many possible reason why this could happen such as this:



Creating a "control" wrapper to isolate how much of the performance difference might be caused by the wedge itself. I have NO idea how Unwinder's tool works, but if it's anything like past attempts from others- it creates a wrapper wedge between application and API- which in turn adds an extra depth of indirection between every API call. This does cause a very measurable drop in performance in itself. A meatless version of the same (same wedge and indirection, but does absolutely nothing) should be benchmarked to remove overhead from the performance results (IF this applies to the implementation of the tool).-Sharkfood from beyond 3d


so just wait and see till/if its confirmed then the flame wars can begin

[Hellbinder]

Nice to see you guys overlooked the other parts of wehat he found. Like Nvidia cheating in Ut 2003 and other games as well. He only mentioned ATi for GT4.

Lets wait and see the Digit life full article before we start calling both companies even in the Cheating department. Dont forget the Recent reviews with Custom no Tweak demos that ATi Wins nearly all of them. Which is not the case with the norm.

I'll let you draw your own conclusions as to what that means.

[ChrisRay]

Just give it more time. I bet both IHV are optimising more than we think.

I would not be surprised if those miracle 9500 Pro drivers were just optimisations...

[StealthHawk]

Originally posted by Hellbinder
Nice to see you guys overlooked the other parts of wehat he found. Like Nvidia cheating in Ut 2003 and other games as well. He only mentioned ATi for GT4.

No one is overlooking anything...both things were posted in the original Unwinder thread here (http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=13217).

Any threads started at Rage3d on this or comments from the fanATIcs at Rage3d about this?

Oh yes, there is.

Discussion (http://www.rage3d.com/board/showthread.php?s=&threadid=33693144).

As can be expected, they simply refuse to believe that there is anything underhanded going on, when I'll bet those same people jumped on nvidia :rolleyes:

Yes there are but plz keep in mind untill we get more detail in terms of how he did it we dont know if it truely a cheat or not there are many possible reason why this could happen such as this:



Creating a "control" wrapper to isolate how much of the performance difference might be caused by the wedge itself. I have NO idea how Unwinder's tool works, but if it's anything like past attempts from others- it creates a wrapper wedge between application and API- which in turn adds an extra depth of indirection between every API call. This does cause a very measurable drop in performance in itself. A meatless version of the same (same wedge and indirection, but does absolutely nothing) should be benchmarked to remove overhead from the performance results (IF this applies to the implementation of the tool).-Sharkfood from beyond 3d


so just wait and see till/if its confirmed then the flame wars can begin

Once again, the performance is not the issue here. I will point you over to this poll (http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=12579) where some 80% of the voters said that ATI's shader instruction shuffling in GT4 of 3dmark03 was cheating because it was an optimization done for a benchmark. Some 69% said they would not consider what ATI did in 3dmark03 cheating if it was done in a game.

Regardless, what ATI has done in 3dmark2001 is much more black and white than what they did in 3dmark03. GT4 of 3dmark2001 is written in PS1.0. Unwinder detected two PS2.0 shaders and one PS1.1 shader in 3dmark2001 GT4. This is hardly a level playing field- in a benchmark.

[ricercar]

ATi has handled their cheats with a lot more class and honesty than nVidia has

Pardon me, but this is like saying that Ronald Regan handled Iran-Contra better than Bill Clinton handled Whitewater or Lewinski.

Cheating is wrong. Saying anything else just obfuscates. Cheating is wrong, no matter who does it or how they react when they're discovered.

[DSC]

Notice how the minute Hellbinder or any fanATIc gets wind of Nvidia cheating in 3dmark03, 3dmark2k1SE or any other benchmark, they posts it in the Nvidia forum. But now that ATI is cheating, they start defending them.... Oh well. :rolleyes:

[ChrisRay]

Whats funny is I remember saying back in an old thread where I was discrediting futuremark.


Saying that the original nature scores we saw before det/cat drivers saw magical increases were more practical performance indicators.


Simply put it. This is no surprise. I'm willing to bet there are "more" ati and Nvidia optimisations in demos ect

[StealthHawk]

Ok, I would like to say that I am wrong when I said thisRegardless, what ATI has done in 3dmark2001 is much more black and white than what they did in 3dmark03. GT4 of 3dmark2001 is written in PS1.0. Unwinder detected two PS2.0 shaders and one PS1.1 shader in 3dmark2001 GT4. This is hardly a level playing field- in a benchmark.

Apparently the PS1.1 and PS2.0 shaders have nothing to do with 3dmark2001. Based on comments at B3D as well as a somewhat poor English translation from Russian I came to that conclusion, but it was erroneous.

Currently I am not sure just what the "shader detections" mean. It is either one of two things, I think.

1) hardcoded shader programs in the driver(which would replace a shader in some game)
2) shaders that the driver looks for, in which case the driver does something.

I am still under the impression that it is #1, just that the detected shaders have nothing to do with 3dmark2001, they are used somewhere else. But Unwinder has said the 1 PS1.1 and 2 PS2.0 shaders detected in the driver have nothing to do with 3dmark2001.

edit: so yeah, I don't know what ATI is doing in GT4 of 3dmark2001, maybe they are doing something as "harmless" as re-ordering shaders. But that's still wrong in a benchmark!

[DMA]

Wow, big surprise. ATi is cheating too :rolleyes:
Why wouldn't they be? Why do everybody think that these Canadians are a bunch of saints? "Ohhh..ATi would never cheat, this must be a misunderstanding!" So tired of that crap..

Kiss my hairy ass fanboys.. :afro:

[goofer456]

Originally posted by DMA


Kiss my hairy ass fanboys.. :afro:
:afraid: :afraid:
No anything but that:barf:

[jbirney]

Well yes ATI did slide some stuff around in 3dmark200x. But nV didn't stop there. From Undwider:


Finally, I'd like to make some comments about the current test results. NVAntiDetector hurts NV performance much more that ATIAntiDetector hurts ATI results. Currently ATIAntiDetector affected performance in both 3DMark2001 and 2003 (performance drop in 2003 is similar to result of installing 330 patch).
NVAntiDetector caused performance drop in a lot of 3DApplications including UT2003, CodeCreatures, AquaMark etc. Performance drop in 3DMark2003 is not comparable to 330 results, results are a way *lower* so it seems like FM missed some detections.

So at least ATI kept its opti-cheats to 3dmarks were nV got a bit more "jiggy" with it :afro: :afro2:

[muzz]

App detection of benches is a cheat( I don't care who it is)......... ATi just got sick of seeing NV get high scores for years doing it, so they did it to keep up.

Sounds familar..........;)

[EMunEeE]

Time to go on a witch hunt....

<gets pitch fork>

I guess I am speechless. Looks like its cheating by both companies. :cool:

Well since everyone's cheating...I don't give a damn now. F*** a benchmark...just let me play my games and DVDs in peace. :p

[saturnotaku]

I think all this goes to show that we all should just shut up and play games. :cool3:

Now where's me Vice City install discs? :bash:

[GlowStick]

Originally posted by jbirney
Well yes ATI did slide some stuff around in 3dmark200x. But nV didn't stop there. From Undwider:




So at least ATI kept its opti-cheats to 3dmarks were nV got a bit more "jiggy" with it :afro: :afro2:

actually ATi didnt just slide stuff around, they completely changed the shaders. Now they refuse to admit to it even though its been widely published.

And also, Every ATi person out there says optmizing for games is 100% competely fine, do you know why? ATi has to do it for every game. Pritty simple.

[Hanners]

Originally posted by GlowStick
actually ATi didnt just slide stuff around, they completely changed the shaders. Now they refuse to admit to it even though its been widely published.

From what I know, ATi did replace the shaders, but with shaders that are functionally identical and have the same output and image quality. I think we've gone over how it's wrong to do this in a benchmark a million times, and ATi have stated it'll be removed in the next drivers - Case closed for 3DMark 2003.

3DMark 2001 is a different matter, that doesn't look as 'innocent' from ATi's point of view. :(


Originally posted by GlowStick
And also, Every ATi person out there says optmizing for games is 100% competely fine, do you know why? ATi has to do it for every game. Pritty simple.

I have no idea where you got that one from - Considering as no evidence of any cheating/optimising in games was found in ATi's drivers, compared to a slew of the stuff in nVidia's, that's one of the more baseless accusations I've heard in some time.

[CapsLock]

I don't really get the rational of this, "lets just stop arguing and play games" position. Its illogical and self defeating. I realize it is a very frustrating and difficult situation, but sticking ones head in the sand doesn't make it go away.

point 1: BENCHMARKS ARE NECESSARY.

How is someone supposed to know how to buy a card without knowing how it performs? How can one keep track of a companies tech development for future purchases or to give advice to another buyer?

Are we just supposed to pick a card at random? Choose by the color of the board? Go with our fanboy loyalty? Choose by whoever has the best PR spin or spends the most marketing?

ponit 2: KNOWING HOW A CARD ACTUALLY PERFORMS IS WHAT ALLOWS US TO ENJOY THE GAMEPLAY AS IT IS SUPPOSED TO BE.

Unfortunately we are in this messy position and it has to be worked out. If we stop caring about honest test results then the consumer will be the loser. (Thats us.)

As for Ati's cheating, I'm sorry they went that route if it turns out to be validated, don't understand why they wouldn't simply reveal suspected cheating instead of stooping to it. (Assuming NV was first?)

Cheating in game benchmarks is the biggest concern here for me. I think its time the web community got it together and stood up to this behaviour, let them know they can't get away with it. Its time for a little reviewer solidarity. They can bully individual sites but they can't do without reviews or the support of the web community.

I'm very glad that some have come out with the custom time demos. This is fighting back and a step in the right direction.

Kudo's to Firing Squad, GamePC and 3DVelocity.

The pen is mightier than the sword !!! Blah Blah Yadda Yadda.

CapsLock

edit: spelling