nVIDIA's Cheating Drivers Have Been Ranked WHQL

[David G.]

WHQL tests both stability and conformance to rasterisation methodology as defined in the API specification and implemented in teh reference rasteriser, its not necessarily perfect at this but it still covers a lot of ground. However, it can not check for application specific optimisations, it would have include teh app itself to do that.

John.

Ok , so it's good that they test for it . The thing is that those "improvements" are not only in 3DMark03 but in any other application . The FPS have gone up in almost every test for the NV3x based cards so the optimisation must be generic and concern the Microsoft tests .

 

[Clashman]

[rant]

I'm actually kind of curious as to why Microsoft would certify these when it won't even run their own damn games! Freelancer crashes and burns with a BSOD with the last several sets of drivers. I have to go back to 41.09.

[/rant]

 

[AzBat]

In the "Cheating and its Implications" thread:

http://www.beyond3d.com/forum/viewtopic.php?p=116951#116951

I mentioned the "GDI Bypass" cheat and how Microsoft tested for it and would not certify drivers that used it. Like others have said in order for Microsoft to do this with timedemo benchmarks they would need to run those tests. I don't know what their test suites look like, but you would think they would run the same games and benchmarks we run and fail the certification if they encountered the cheating. But again Microsoft is probably only concerned with how drivers affect the OS and not specific applications.

Tommy McClain

 

[JohnH]

The problem is that the tests take to long to run even now, making them even more extensive is just going to make that worse...

John.

 

[MrGaribaldi]

The thing is that those "improvements" are not only in 3DMark03 but in any other application . The FPS have gone up in almost every test for the NV3x based cards so the optimisation must be generic and concern the Microsoft tests .

Well, couldn't some of those improvements come from actual proper optimization, and not just from cheating?

Unless I've misread this board (which is of course an option since I'm not to knowledgeable about 3d tech/drivers), inserting the form of clipping/culling seen in 3dmark requires some work for each application and isn't too hard to notice once you look for it, thus making it unlikely it's happening in every app...


I'm not really trying to defend Nvidia, but rather the notion that not all of their driver team(s) is working on cheats...

 

[David G.]

The thing is that those "improvements" are not only in 3DMark03 but in any other application . The FPS have gone up in almost every test for the NV3x based cards so the optimisation must be generic and concern the Microsoft tests .

Well, couldn't some of those improvements come from actual proper optimization, and not just from cheating?

Of course .

The thing is that the increase delta is almost constant in all the application range ( 33% ) which leads me to belive the omissions ocur everytime .

And the thing that proves the constancy is 3dmark03 itself , even if in some tests the increase is close to 50% , the over all incheatement err... improvement is still around 30% .