3DMark03 Patch 330 Released

[TS | Thomas]

We have now established that NVIDIA’s Detonator FX drivers contain certain detection mechanisms that cause an artificially high score when using 3DMark 2003. We have just published a patch 330 for 3DMark03 that defeats the detection mechanisms in the drivers & provides correct results.

Download the 3DMark03 Patch 330 here:
http://www.futuremark.com/download/?3dmark03patch330.shtml
A detailed audit report of the drivers & 3DMark03 is available at:
http://www.futuremark.com/companyinfo/3dmark03_audit_report.pdf

I'd highly recommend everyone read Futuremark's report, which details the Driver cheats NVIDIA implemented. I'd be interested to hear what NVIDIA now it's been proven these are far from Driver "bugs". Course, the main problem is there's absolutely no punishment for NVIDIA for misrepresenting their Graphics Cards performance with those Drivers.

 

[negroplasty]

Let's all notice what else that little PDF says. Notice on - okay, I was going to give a page number for this but my Acrobat is messed up. Anyway, to the point, if you search carefully enough (around page 4), you will notice Fururemark saying how ATi also "cheats" in 3D Mark. Just want to hear some people's opinion on that, thanks to MrGaribaldi for this info.

 

[DigitAlex]

i'd like some other people not trying to turn every information against some companies there are facts, so why just go over again with nv vs ati ??

yes, we all noticed they were saying that THE RESULTS WERE DIFFERENT and that THEY WILL INVESTIGATE THAT FURTHER, not that somebody else was cheating. by the way, on the other thread there was a link to a page about another company cheating in drivers (not ATI nor NV) ...

 

[negroplasty]

It was not my intention to start a discussion about NVIDIA vs. ATi, nor did MrG have anything to do with this, he solely pointed out this interesting fact/theory. Which is what I wanted to do, I wanted people to see that NVIDIA is not ALL BAD, and other companies DO cheat, or whatever you want to call this. Only one of them was smarter than the other, ATi in this case . Once again, I only wanted to point out that cheating is not all it is made out to be, and to be honest, it isn't a very serious event. Think of it from NVIDIA's side of the story, they think they are getting ripped off by Futuremark because they left the BETA program and that the benchmarks are tilted towards ATi's benefit; would you not cheat, or optimize your drivers if you were in the same situation? And from ATi's side, they see NVIDIA cheat, which worries them, so they respond by doing the same thing and so on and so forth.

 

[acidosmosis]

I don't use benchmarks anyway so I could care less. I want REAL tests of my hardware, not some benchmark gimmick. The only thing they are useful for is seeing some neat animation. Anyone relying on benchmark's is for lack of better words, and the flat out truth... an embacil.

 

[mrslippyfist]

Originally posted by timmoore
I wanted people to see that NVIDIA is not ALL BAD, and other companies DO cheat, or whatever you want to call this.

Of course it makes Nvidia all bad, just cos another company has also cheated doesn't make it alright, in fact with Ati doing the same, although apparently to a lesser degree, makes a mockery of the benchmark. Effectivly they have been falsely advertising their cards, claiming higher proformance than in reality, encourging people to buy their product, and not living up to expectations. I don't want to spark some big argument about which card maker is better, but if card makers feel that they are allowed to fiddle the benchmarks, it means there will be no easy way to judge one card over another, it is sad that both companies have felt that market pressure requires them to do this.
Btw I have a gf4 ti4400, so am not an Ati fanboy, bought the best card I could for the money (£130 on release day cos of a mistake with the pricing)

 

[Unregistered]

<insert any computer company> fanboys should get a life.

 

[Rick]

... The things companies do for money. :\

 

[MrGaribaldi]

Originally posted by mrslippyfist
Of course it makes Nvidia all bad, just cos another company has also cheated doesn't make it alright, in fact with Ati doing the same, although apparently to a lesser degree, makes a mockery of the benchmark.

100% agree with the first statement! Nvidia has done a bad thing!

As for ATI, have they cheated or not?
Olefarte started a thread here, which has what claims to be ATI's reply to the allegations...

If it really was from ATI, and what they're saying it's true, it'll be a bit hard to say if it's a cheat or not... (Bear with me for a second before you attack me)
It's a cheat beacuse:
1. It recognizes the benchmark.
2. It changes the shadercode in GT 4

It's not a cheat because:
1. It gives the same visual effects as the original code
2. It is running at full precision
3. It is just shuffeling the original code so it's more efficient when going through the pipeline... (Not really changing the code, just changing which effect is sendt first)

So I'll want to wait on FM (to know what they define as cheating with 3dmark) before I making a conclusion...

And no, I'm not treating ATI any differently than I would nvidia in this case...
But what nvidia did (with Det FX 44.03) was obviously a cheat (from what ET had found), whereas this has the potential of being a cheat and not being a cheat depending on how FM sees it...


timmore> So you mean that if USA cheated in the olympics (taking performance enhancing drugs), it wouldn't be a bad thing, since China had done the same?
Or that Enron's financial cheating wasn't a bad thing, since some other company has done the same at one time or another?

In my opinion Nvidia has done a bad thing and should pay the price for it... If it's proven that ATI has done a bad thing, then they too should pay for it...

I get aggrieved when people are saying, "it doesn't really matter if someone cheats, since others are doing it too..."

Don't you want to be able to trust a review when you read it?
Would you really like it if the only way to see how a card would perform was to buy it and test it yoursefl?

Personally I'd like to be able to log onto a site, read a review and not have to wonder if those numbers where real, or if the card had some cheating drivers...

We can never get away from the possibility of the reviews being skewed towards one IHV or another, but we should be able to know and take the numbers for what they really are!

I seriously hope that Nvidia landed a deal with an OEM because of the inflated numbers, and now have to bear the brunt of that OEM's anger... That should give them a hint that they shouldn't do it again...

.02$

 

[negroplasty]

I apologize for aggrieving you MrG, you are right. NVIDIA has done something wrong, but this does not necessarily make them bad, nor does it ATi if they have indeed cheated. I think that NVIDIA should pay the price for their actions as they were both treacherous and deceitful. As for your metaphor, I believe I see your point, and agree with you on the matter. But once again, if you were in NVIDIA's position, and you thought you had been ripped off by Futuremark; would you not do the same thing or something very similar? I know that I would, cheating is wrong, which is why they should have just made it so that 3DMark couldn't run on NVIDIA cards, which I expect to be the next step NVIDIA takes in their next Detonator release.

This post pretty much sums up my feelings on this subject.

 

[MrGaribaldi]

Timmore> Thank you

You are handling this very maturely!

If I were in Nvidia's shoes/were supposed to take out FM...
Well, I'd either disable support for 3dmark in the drivers (as I've said before), or I'd release a "special" driver set, called something like 3dmark03 killer, which had all these cheats in them, BUT with a very big warning telling people that this driver set had only one intention, and that was to show them how easy it is to cheat in 3dmark...


As for Nvidia and/or ATI being bad... I take your point...
But it'll be interesting to see what you think in a week or two

 

[negroplasty]

Thanks a lot MrGaribaldi; I am also looking forward to see what I think in the near future. As somebody once said, we should enjoy the competition while it lasts, as we solely benefit from it. Which is completely true, NVIDIA and ATi competing is necessary to stop one of them rising prices or reducing quality of their products. With this competing going on, we benefit by lower prices, higher quality and more products being released.

I can only begin to explain the excitement I experience at the release of a new product, or new drivers for that matter. I think ATi and NVIDIA need each other to survive, I only favor ones products as they were the first company I have been with and they have never given me any trouble, which the same cannot be said about ATi (I have had trouble with their hardware in the past) .

 

[DigitAlex]

Originally posted by timmoore As somebody once said,

Well I remember saying something about that, about competition almost always have been good in the past (usa vs ussr in space, archeologists with dinosaurs' bones, etc)