Nvidia just got caught cheating AGAIN

[denta80]

man, what's wrong with nvidia?
seems like all they have done lately (since they came up with FX series) is hyping their products and cheating.
i hope they get their acts together soon, or they are going down fast.




Nvidia caught cheating AGAIN
http://www.theinquirer.net/?article=15502

Radeon X800PRO will beat NV40 Ultra
http://www.theinquirer.net/?article=15406

ATi is offering native pci express cards where nvidia's cards are "bridged"
http://www.ati.com/products/PCIexpress/index.html

The Inquirer's statements regarding pixel shader revisions can not be verified by any industry source, thus I will remain silent concerning those claims until further independent testing can be performed.

Anyway, I highly recommend everyone read the original article regarding the accusations of cheating.

http://www.driverheaven.net/articles/driverIQ/

DH has capitalized on the publicity of finding a mipmapping bug in non-WHQL beta drivers. Am I supposed to be impressed with this finding?

I have found similar problems during the review process for graphics chipsets. I usually contact the manufacturer's driver development team first, instead of rushing out to publish an editorial claiming a company is somehow "cheating" the consumers out of a partial-texture of fine detail resolution.

Simply put, nearly every manufacturer has failed stringent IQ testing (especially in FutureMark tests) at some point, including ATI. The majority of the time it is a simple driver bug that slipped passed the manufacturer's quality assurance team. I would tend to believe this is the case this time around, especially considering the beta nature of the NV40 driver architecture.

Only when nVidia can not or refuses to fix the issue will I then agree "cheating" is taking place in this situation.

Robert Richmond

[phrozen_dreamz]

... a little off topic but, Rob I am geussing that you're no high schooler. I have been reading some of your posts as of late, and your proffesionalism is something of high quality. What do you do for work?

http://www.techimo.com/about.html

I have been a corporate journalist within the x86 hardware sector since 1998. While my employment title has changed over the years, I have a journalistic background in processor design and fabrication, with the occasional stint in graphics and interconnect subsystems.

I have not had the chance to develop much technical content for TechIMO due to other ongoing responsibilities within the company, such as ResellerRatings and public relations.

Robert Richmond

[phrozen_dreamz]

I have been a corporate journalist within the x86 hardware sector since 1998. While my employment title has changed over the years, I have a journalistic background in processor design and fabrication, with the occasional stint in graphics and interconnect subsystems.

Robert Richmond [/B]

See what I mean?
Well at least you post quality posts.
Unlike some members

[Bizkitkid2001]

I still don't know why people call it cheating. Cheating is to harsh of a word. Getting companies to make a product run faster on your product is what nvidia is doing. There not cheating by making it seem there card is running faster than it is suppose to. Also, who knows how fast the X800 will be? Only ATI does, so that article is irrelavent to this discussion.

For me, I go for the cheaper card that offers the best performance. I don't care what they have to do to get the card to run faster/better, if it runs faster/better, why all the complaint? Especially if you can't tell with the naked eye.

[pinipig523]

cheating is cheating - ur saying that it is ok for a runner to have a two leg head start in a 100m dash.. i mean, after all - all we're after is the faster time right?

if u condone cheating at this level, where will it stop?

[slinky]

This is coming from the inquirer... that should have given it away as to the validity of the claim.

if u condone cheating at this level, where will it stop?

How are we to know this "cheat" is nothing more than a simple error in nVidia's mipmapping routine? Am I to believe it is a "cheat" because a couple of hardware sites opted to run this story for whatever reasons? You can witness that this article did not appear in TechIMO's front-page news for Friday.

Such a simple mipmapping error has happened before on many occasions dating all the way back to the earliest 3dfx 3D cards, why not again? All it takes it a slight deviation in a complex mathematical calculation, or a numerical truncation too early in the pipeline, or a slight rounding problem, or.... you get the point.

ATI has also been found (even admitted) optimizing drivers for FutureMark benchmarks (among other popular 3D programs) over the past months, but no media sources seem to ever discuss those "cheats." Sadly, this is because most media review sites have to directly negotiate with ATI to obtain review hardware. In contrast, these same sites can often bypass nVidia and go straight to board vendors to acquire nVidia review products. I hope everyone sees a distinct difference in those two scenarios from a media standpoint.

Even worse, companies like Trident have used driver-level auto-detection routines to determine when the 3DMark executable was being loaded to enact a set of render "optimizations" that make nVidia’s "cheats" appear to be legitimate features in comparison. Forget about mipmapped textures, how about whole missing frames?!

Robert Richmond

[Xtreeme]

When the drivers go official OR Nvidia tried tries to use it as bench refrence to competition THEN its cheating. Till then-its in development. If you cant buy one-use one or sell one, how does it matter yet? Unless as I said above happend.