Is NVIDIA cheating to sell graphic cards?

[Slappi]

Ok so I read the article. Some great investigative reporting there let me tell yah. First those numnuts don't even test older drivers to compare, then they don't even try to recreate the same thing in any games or if they do can't do it and don't say. The whole article stinks of bias without any reason except for their shotty "investigation". Wow before I made some HUGE assumptions I would get a few more facts and talk to a few more people that know more first.

3Dmark03 plain sucks. That is all there is to it. The whole benchmark is ghey. If I want benches I do em myself with a GAME!! That POS isn't worth the space on my HD it takes up. I know it is said a lot but I will say it again ....... I don't play 3Dmark!!!!

I agree with the other poster ...... NV should quit screwing with 3Dmark and let it die which would be a great thing IMO!!!

[nVidi0t]

Originally posted by ZoinKs!
You're right. Higher cpu speed has little impact on 3dmark03. And that's a good thing.

Q: Was 3dmark created to test the performance of the cpu? Or to test the performance of the vidcard?
A: It is meant to test the performance of the vidcard.

These two statements are true:
1. Changing cpu has low impact on 3dmark03 results.
2. Changing vidcards has dramatic effect on 3dmark03 results.

Conclusions: 3dmark03 is an excellent method of measuring vidcards. It is a poor way of measuring cpu's. Therefore, 3dmark succeeds as a vidcard benchmark.

If you want to measure overall system performance, then you need to use a different benchmark. If you want to measure how well a vidcard performs with dx 9 pixel and vertex shaders, 3dmark03 is an excellent tool.



btw, nv40: PS 1.4 is directX 9 spec...if the FX can't handle dx 9 specifications, whose fault is that?

I do not believe AT ALL that 3dmark03 was intended to be a vidcard benchmark.

Why the Audio tests? Why the CPU tests?

Why in the Orbb results viewer does it list entire system specs? It is a benchmark which tests the 3d performance of your entire system.

And for that it's a terrible benchmark.

[zakelwe]

3dmark does have it's benefits Slappi due to it's very large database of different systems.

It's interesting to note that only the DX9 GT4 game shows the problem, why does GT2 and GT3 not show it ?

You get more points for GT3 realtively and so there would have been the place to "cheat" . Then my GF4 would have benefitted as well of course

It's terrible for nVidia to exclude us GF4 guys !

If the released Det 50's / FX don't show this problem and are as fast then ExtremeTech might find themselves in serious legal problems.

Of course, I use LOD adjustments when I run 3dmark so I can't really have an unbiased comment on the matter.

Regards

Andy

[zakelwe]

Originally posted by nVidi0t
I do not believe AT ALL that 3dmark03 was intended to be a vidcard benchmark.

Why the Audio tests? Why the CPU tests?

Why in the Orbb results viewer does it list entire system specs? It is a benchmark which tests the 3d performance of your entire system.

And for that it's a terrible benchmark.

No, the video section of the benchmark just tests your video card ( at least compared to 2001 ) .. that's what he's saying.

The games in 2001 test your entire system and are more relevant to current games.

Regards

Andy

[Hanners]

The Iraqi Information Minister would be proud of some of you guys.

[Nemesis77]

Originally posted by nVidi0t I do not believe AT ALL that 3dmark03 was intended to be a vidcard benchmark.

Why the Audio tests? Why the CPU tests?

So the users could see how the CPU and sound-card affects performance? But to my knowledge, the 3DMark-score is calculated from the game-test, NOT from the sound-tests. The other tests are just extra.

Why in the Orbb results viewer does it list entire system specs? It is a benchmark which tests the 3d performance of your entire system.

Why shouldn't it show all the specs?

And for that it's a terrible benchmark.

You forgot to mention your real reason for believing that: "Because NVIDIA sez so!"

[vampireuk]

Right I'm getting sick and tired of the childish bitching and nit picking about the 5900 because some people just do not like NVIDIA. How about I make a picture of some NVIDIA engineers eating a baby, you could do with something new to cry about

[nVidi0t]

Originally posted by Nemesis77
So the users could see how the CPU and sound-card affects performance? But to my knowledge, the 3DMark-score is calculated from the game-test, NOT from the sound-tests. The other tests are just extra.



Why shouldn't it show all the specs?



You forgot to mention your real reason for believing that: "Because NVIDIA sez so!"

No, because I have tested it, and major differences to system spec, but not changing the display card does not affect performance in the benchmark, yet in games, the average FPS is 50% faster.

It shows all the specs, and allows you to compare your results to other users with the same CPU, obviously it's a system comparison.

The sound tests do impact score. And the CPU tests should aswell, yet trying an exact system spec with another CPU almost 50% faster, the 3dmark scores remain the same.

It's a bad benchmark. Game engine benchmarks are the only real way of determining performance WHICH MATTERS to the customer, they are going to be playing games, not benchmarks. (unless they are like a few peeps here.. lol)

[Nemesis77]

Originally posted by Slappi Ok so I read the article. Some great investigative reporting there let me tell yah. First those numnuts don't even test older drivers to compare, then they don't even try to recreate the same thing in any games

How could they recreate it in games, since this "bug" only affect benchmarks where the camera goes along a fixed path? Luckily in 3DMark you CAN move that camera around, and that exposes this cheat.

The whole article stinks of bias without any reason except for their shotty "investigation". Wow before I made some HUGE assumptions I would get a few more facts and talk to a few more people that know more first.

And your comments stink of blind "NVIDIA can do no wrong! NVIDIA never cheats!"-bias. Seriously, is the Iraqi Information Minister your mentor?

3Dmark03 plain sucks.

Yeah, because "NVIDIA told me so!". How exactly does this thing show that it sucks? Quite the contrary, it shows why it doesn't suck: cheated get caught.

That is all there is to it. The whole benchmark is ghey.

Funny, nvidiots didn't have any problems with 3DMark when it was still officially approved by NV

If I want benches I do em myself with a GAME!!

And this cheat could be applied to timedemos in games as well.

That POS isn't worth the space on my HD it takes up.

My my, aren't we defensive ?

I agree with the other poster ...... NV should quit screwing with 3Dmark and let it die which would be a great thing IMO!!!

Yeah, because all benchmarks must be "officially approved by NVIDIA" and they must show NV hardware in best possible light. If they don't do that, the benchmark in question must die.

[Humus]

So 1 in 3 of you don't think nVidia cheated? And I was annoyed over how many fanATIcs who denied that ATI cheated with quack ...

[vampireuk]

24 people have voted no, it hardly represents the entire online community does it.

[AnteP]

Originally posted by nVidi0t
No, because I have tested it, and major differences to system spec, but not changing the display card does not affect performance in the benchmark, yet in games, the average FPS is 50% faster.

It shows all the specs, and allows you to compare your results to other users with the same CPU, obviously it's a system comparison.

The sound tests do impact score. And the CPU tests should aswell, yet trying an exact system spec with another CPU almost 50% faster, the 3dmark scores remain the same.

It's a bad benchmark. Game engine benchmarks are the only real way of determining performance WHICH MATTERS to the customer, they are going to be playing games, not benchmarks. (unless they are like a few peeps here.. lol)

You my dear sir is a liar.

No score except the ones from game Test 1 through 4 has an impact on the final score.

3Dmark03 was DESIGNED to put the load on the videocard. You know why?
Because that's the future of shading intense games.

Here's what Futuremark had to say about even before it was available for download:
http://www.nordichardware.se/artikla...sh/Futuremark/

[NH]: There has been some complaints that the CPU, RAM and MOBO influenced the scores too much, is this any different in the 03 version?
[FM]: As we have focused even more heavily on the Shaders in 3DMark®03, the graphics card is more of a bottleneck than it was in 2000 and 2001. 3DMark®03 comes with a separate CPU test for more in-depth CPU benchmarking.

Iraqi Information Minister indeed.