I told you so - ATI cheats also

[AmdInside]

During the technical presentation, there was a surprise of sorts. NVIDIA, along with the assistance of Mark Rein from Epic Games, pointed out a rendering flaw in ATI's R360. The ATI hardware did not seem to render all of the detail textures. This was a huge surprise to the audience, as none of us had seen the flaw in ATI hardware before.

There was a second example in the AquaMark 3.0 benchmark in the Massive Overdraw scene where NVIDIA pointed out another flaw in ATI's rendering strategy. The scene should look like a washed out bright spot, but the 'Red' company's render of the same frame looked like several textures were being left out.

There were also quite a few pointed barbs at ATI, oops, I mean the 'Red' company. All of the developers present, some seemingly acting out some sort of strained confessional, were pointing out that sure, the ATI cards were indeed faster, but they also weren't drawing everything either, like the detail textures and alpha values. These kinds of image quality issues are hard to prove, so take it with a grain of salt.

Apparently NVIDIA has known about these particular ATI issues for some months now, but as a general practice, doesn't write white papers about its competitor's hardware.

http://www.guru3d.com/article/article/91/1

 

[NOX]

C'mon you know no one wants to hear about this stuff unless it's Nvidia cheating!

 

[nimo]

Coming from a competitor with a record of "tweaking" the truth them selves the credibility is some what lacking

 

[arcenite]

oh man... i just bought an ati card cause nvidia is cheating.. now ati is cheating... omg i'm gonna have to get a MATROX!

Bill

 

[rbV5]

I can see a "huge" difference in the images on page 2, the Radeon shots look completely out of focus...not sharp at all. You can't even read the text at all in the upper left hand corner, yet on the top shot the text is clearly visable.

 

[JBT]

Originally posted by: rbV5
I can see a "huge" difference in the images on page 2, the Radeon shots look completely out of focus...not sharp at all. You can't even read the text at all in the upper left hand corner, yet on the top shot the text is clearly visable.

I see that too. While the 9800 shot is clearly out of focus it is still possible that the image quality isn't as good to bad you have to blow it up to fit and enite wall to see it lol.

 

[LikeLinus]

Originally posted by: rbV5
I can see a "huge" difference in the images on page 2, the Radeon shots look completely out of focus...not sharp at all. You can't even read the text at all in the upper left hand corner, yet on the top shot the text is clearly visable.

Looks like someone at nVidia accidentally pushed the focus button on the projector during their presentation.

The whole frame is out of focus. Yeah I believe nVidia on this one...lol

 

[GTaudiophile]

Talk about grabbing at straws. If nVidia had faith in their hardware, there would be no need to point fingers like this.

 

[DaveSimmons]

Probably true, but are they intentional like nV's last half-dozen cheats have been, or are they just true bugs like both nV and ATI have always had in some games?

And will ATI lie about it like nV has done repeatedly?

Beats me, I'm still a happy GeForce3 owner (back when nV made good cards & drivers) waiting to upgrade to either ATI or nV after the dust settles a little.

 

[Sid59]

bah, it's all scheme to 'level' the hardware field.

 

[Nebor]

Who cares?

 

[Megatomic]

Originally posted by: GTaudiophile
Talk about grabbing at straws. If nVidia had faith in their hardware, there would be no need to point fingers like this.

What comes around goes around. Karma is a b!tch sometimes...

 

[Dogmeat]

Can I use the ATI cheat codes with my NVIDIA card??

 

[Pocatello]

Everybody cheats. I've noticed that my Radeon 9800 pro benchmark scores tend to jump really high then suddenly drop really fast. I wonder if this is the reason why on average, the Radeon 9800 does so "well".

 

[Matthias99]

The *photo* of the ATI screenshot on page 2 of that article is badly out of focus (probably not using a tripod). Thus, it's not really possible to see how the screenshots compare unless NVIDIA releases their presentation to the public.

<dons fireproof suit> The Tom's Hardware 5700 review today also touched on this, and had screenshots from AM3 (though one of the full-size comparison images isn't working -- it links to the wrong picture). To me, the explosions look pretty damn similar, even side-by-side -- though the difference stands out more in their back-and-forth animated GIF. Apparently, they're not drawing some layered transparent textures, although it's unclear if this is a bug, feature, or benchmarking-only optimization, since this is the only place anyone seems to have seen it. My guess (and it's only a guess) is that this is an across-the-board optimization where they're not drawing overlayed transparent textures that appear to have little to no impact on the scene (although in this case they do, since there's lots of them). There was also a discrepancy with a couple of the particle effects, but that hadn't been mentioned by NVIDIA. They couldn't find anything wrong in UT2K3 or Halo, but they also didn't have screen grabs in the exact same places.

As I recall, didn't ATI's scores in 3DMark03 also drop a bit when Futuremark put out that patch to disable the drivers from detecting that the benchmark was running? Everybody optimizes for specific applications -- it's just that ATI has (mostly) done it so far using methods that don't noticeably impact image quality.

 

[AmdInside]

More info from Toms Hardware:

Earlier this year, NVIDIA drew a lot of flak over a number of questionable optimizations in their drivers for 3DMark 2003 - optimizations which could definitely be called cheats. It was in response to this that NVIDIA came up with the guidelines pertaining to performance optimizations mentioned above. Laudably, we haven't been able to make out any "violations" of the new rules in the new driver v52.16 - at least so far.

ATi, on the other hand, repeatedly claimed that their drivers implemented no illegitimate optimizations. The accusations leveled against ATi at NVIDIA's Editors' Day two days ago thus become that much more serious. Epic's Mark Rein confirmed that in some cases, high-res detail textures were not displayed in some areas by ATi´s drivers and that standard, lower-res textures are used instead. Randy Pitchford of the Halo development team also mentioned that there were optimizations present in ATi's drivers which are detrimental to Halo's image quality. However, Randy didn't want to go into more detail here. Finally, Massive's new DX9 benchmark, AquaMark 3, also displayed some irregularities of ATi drivers in the overdraw test.

Here is a list of accusations we were able to confirm:

http://www.tomshardware.com/graphic/20031023/nvidia-nv38-nv36-19.html#questionable_optimizations_in_atis_drivers

 

[SithSolo1]

Originally posted by: AmdInside
More info from Toms Hardware:

Earlier this year, NVIDIA drew a lot of flak over a number of questionable optimizations in their drivers for 3DMark 2003 - optimizations which could definitely be called cheats. It was in response to this that NVIDIA came up with the guidelines pertaining to performance optimizations mentioned above. Laudably, we haven't been able to make out any "violations" of the new rules in the new driver v52.16 - at least so far.

ATi, on the other hand, repeatedly claimed that their drivers implemented no illegitimate optimizations. The accusations leveled against ATi at NVIDIA's Editors' Day two days ago thus become that much more serious. Epic's Mark Rein confirmed that in some cases, high-res detail textures were not displayed in some areas by ATi´s drivers and that standard, lower-res textures are used instead. Randy Pitchford of the Halo development team also mentioned that there were optimizations present in ATi's drivers which are detrimental to Halo's image quality. However, Randy didn't want to go into more detail here. Finally, Massive's new DX9 benchmark, AquaMark 3, also displayed some irregularities of ATi drivers in the overdraw test.

Here is a list of accusations we were able to confirm:

http://www.tomshardware.com/graphic/20031023/nvidia-nv38-nv36-19.html#questionable_optimizations_in_atis_drivers

Thats about like saying more info from the homeless guy I found outside my office.

 

[Sunner]

Originally posted by: GTaudiophile
Talk about grabbing at straws. If nVidia had faith in their hardware, there would be no need to point fingers like this.

Somehow I get the feeling that isn't quite what you'd be saying if the tables were turned.

 

[EeyoreX]

Frankly I find two good quotes in this thread:

Who cares?

and

Everybody cheats.

If you believe for a moment that there is a hardware company that doesn't use certain benchmarks to their advantage via "tweaks" or whatever, I think you are just fooling yourself. Every hardware manufacturer wants to look their best/better than everyone else so they either use benchmarks that are to their advantage or "tweak" settings. This is the reason that I, and many others, do not put too much (any) stock in synthetic benchmarks. I certainly do not base any product purchase on a benchmark.

\Dan

 

[Medicated858]

That's why I only put SIS Xabres in all my computers!!

 

[ViRGE]

Originally posted by: aRCeNiTe
oh man... i just bought an ati card cause nvidia is cheating.. now ati is cheating... omg i'm gonna have to get a MATROX!

Bill

Didn't you hear? Matrox is the worst cheater of them all. The chip on the Parhelia isn't real; all the rendering is really done via software.

 

[MDE]

Now that Nvidia has this round of publicity over, where's my updated nForce drivers?!

 

[wetcat007]

Originally posted by: Dogmeat
Can I use the ATI cheat codes with my NVIDIA card??

LMAO

 

[PraetorianGuards]

Originally posted by: wetcat007

Originally posted by: Dogmeat
Can I use the ATI cheat codes with my NVIDIA card??

LMAO

haha great.

Seriously though, this was released at an nVidia bought and paid for event, so until more independent sites (not biased THG) come out with proof then I'll wait.

 

[arcenite]

Originally posted by: ViRGE

Originally posted by: aRCeNiTe
oh man... i just bought an ati card cause nvidia is cheating.. now ati is cheating... omg i'm gonna have to get a MATROX!

Bill

Didn't you hear? Matrox is the worst cheater of them all. The chip on the Parhelia isn't real; all the rendering is really done via software.

Oh man... Trident, Here I come!!

Bill