Nvidia - caught cheating again?? New Aquamark

[Supa]

From Tom's Hardware

Link


"Bug" or "Cheat"? - NVIDIA drivers don't render the correct image

Before we show the performance results of actual cards in the benchmark, we need to take a look at the image quality of the cards and different driver versions first.

The image comparisons we made between the ATI Radeon 9800 and NVIDIA FX5900 cards quickly made it clear that something is wrong with the NVIDIA driver. The lighting in terrain and objects of v45.23, for example, do not work.

NVIDIA explains it's a bug but how this bug got into the NVIDIA drivers is unclear. The fact is that the old driver version v44.03 renders a fully correct image but its performance is very low. ATI and NVIDIA both had worked with beta versions of AquaMark 3 for months, and we even had an early beta version of AM in April. So it's hard to understand how this could happen. Even the new Detonator v51.75 beta driver that was sent to the press last week by NVIDIA, and which was claimed to be the driver to use for DX9 benchmarking like Half-Life 2 and AquaMark 3, renders an incorrect image.

We also tested an older ATI driver (Catalyst 3.4), which was released at the same time then NVIDIA´s v44.03. We found the following problems the drivers posed:

* NVIDIA v44.03: Renders a correct image. Faultless anisotropic filtering;
* v45.23: Deterioration of quality. Light on terrain is not displayed. Light on objects is not displayed (sometimes). Shadows are not displayed;
* NVIDIA v51.75: Deterioration of quality. Light on objects is not displayed. Very low anisotropic filtering quality.

The performance numbers change dramatically. From v44.03 to v45.23, the AM3 Triscores climbs 71% and 82% to 51.75 ....


---

 

[rocadelpunk]

dun dun dun

 

[Corporate Thug]

OMG, Nvidia cheating? say it aint so!

det50 is garbage. it renders like crap to churn out more frames

 

[ethebubbeth]

:Q omg omg omg.... nvidia is trading IQ for performance!!!!!!!!!! NO WAY!!!!!!!!!!!!!!!!!!

 

[Supa]

I own Geforce4 4200, just went back to the old 44.03 driver. Thought 45.23 is the version without cheat, guess I was wrong to believe ...

Nvidia's New Driver Optimization Guidelines


---

 

[KGB1]

Originally posted by: Supa
I own Geforce4 4200, just went back to the old 44.03 driver. Thought 45.23 is the version without cheat, guess I was wrong to believe ...


---

HEhe...I went from 45.23----->41.09

Its working like before, I don't really see a difference and the 45.23 problems can probably be mainly caused by GF/FX hardware problems and not comparable to Ti4200 problems.

 

[Vonkhan]

... and the nVIDIOTS claim tht ATi has buggy drivers

 

[Supa]

On 1024 x 768 no AA, no AF, 5900 Ultra's mark is about the same as 9600 Pro's with 44.03, but Jump to 9800 Pro's level with 45.23 and 51.75 drivers.

And on 1600 x 1200 no AA, no AF, 5900 Ultra's mark more than doubled from 44.03 to 45.23 and 51.75!!! Didn't know driver "optimization" can double your performance.

- "We colored the scores with the NVIDIA drivers v45.23 and v51.75 because they do not render the correct image (see quality comparison). The only driver rendering a correct image is v44.03 from May 2003, which shows a very low performance compared with the ATI cards. "


---

 

[McArra]

Just another one. But not everything seems cheating, there also are bugs.... one bug.

 

[Johnbear007]

Originally posted by: Vonkhan
... and the nVIDIOTS claim tht ATi has buggy drivers

 

[Genx87]

Tell me about it.

IQ

BTW this was taken on a 9800.

I would wait till the final release of the drivers occurs. Right now you are getting all hyped up over beta drivers.

BTW the article mentions how ATI is having a problem applying AF to certain angles.
Overall I wasnt terribly impressed with either cards ability to render the images correctly.

 

[Supa]

Don't think 45.23 is beta, and the point is the underlying intention, which is a far cry from Nvidia's own guidelines.


---

 

[jiffylube1024]

Originally posted by: Vonkhan
... and the nVIDIOTS claim tht ATi has buggy drivers

There is a difference. nVidia cheats benchmarks by lowering IQ and turning off some effects. ATI has bugs in some games that are not deliberate. However, ATI has done an excellent job in improving the Cats, making them almost problem free in pretty much every game I run. IMO (before this whole nVidia fiasco) nVidia had a little bit tighter drivers, but now I just don't know what to think. They have been caught with their hand in the cookie jar so many times, but they keep going for it. It makes absolutely no sense to me...

 

[Lonyo]

Hod does nVidia's driver team do it? In every new set of drivers, they seem able to pull more and more performance out of the bag. I don't really remember this as much with the Ti4xxx cards, at a time when nVidia had not much competition, and yet now they have to content with ATi they seem able to continually increase performance up to 100% with driver releases. Magic.

Maybe their hardware is so much more powerful than we ever imagined, and they're just slowly releasing its power with new driver releases.

(User of a Ti4400 from ~April last year, and haven't seen huge jumps in performance from new drivers)

 

[edmundoab]

Originally posted by: shady06
OMG, Nvidia cheating? say it aint so!

det50 is garbage. it renders like crap to churn out more frames

damn, if this is the kinda image NVIDIA is giving to the public every now and then....
the sales is going to ski way down...

 

[Viper96720]

Wonder how long Nvidia actually has actually done it? I believe just recently but what if they've done the same things before. But the video cards didn't have as much capability as
the new ones. Now since the IQ is better you notice the differences while before you couldn't. Also IQ didn't seem to be scrutinized as much in older reviews. How they got jumps in performance from each previous detonator release.

 

[Rogodin2]

From a review at beyond3d and interviews with devs there is only one way to make games run faster with driver ops-that is to decrease IQ-period.

ATI does it to a lesser extent than nvidia but it's just the way things are-"To be is to be Rendered."

Plontinus;o

rogo

 

[DaveSimmons]

From a review at beyond3d and interviews with devs there is only one way to make games run faster with driver ops-that is to decrease IQ-period.

mostly true for game-specific optimizations, but not true for overall performance since there you can definitely profile, tune, bugfix, and do real (non-cheating) optimizations like trading memory use for speed.

Even per-game you can have optimzations like ATI's 3DMark03 shader cheat where they built a faster shader than Futuremark's but with the same image quality, in other words if the driver devs write faster replacement code to make up for slow code from the game developers.