Why Nvidia 42.86's score better ? ,..see by yourself.

[TimisoaraKill]

Why we get a better score with the magical 42.86 'optimised' 3Dmark 2003 , because they cut explosions in the demo and the fire from machineguns .
This could be or a cheat or a driver issue , look like they optimised a bit to much .
got this from here : Linky




/////////////////////////



This is how it sould look :

[Gabrielx]

Most interesting.

It could be a driver-issue of course, but it wouldn't surprise me a bit if it wasn't...

[Paladyr]

I don't think explosions just magically disappear :-P and there is no way that they didn't try out 3dmark2003 before releasing these.

[omeo]

If this is true. I now have less respect for Nvidia than ever before. Ati cheated with "quack" a year back. Nvidia is doing the same, while simeltaneously claiming that the benchmarks are unimportant?

[s1203372]

Is it possible that the fire/explosions use DX9 effects which the GF4's can't create? Do older driver revisions show these effects? Also, does anyone have the same scene so we can evaluate apples vs. apples? Even better, does anyone have an 8500 vs. GF3/4 shot of the same scene?

It will be very interesting to see if Nvidia pulled a quack.

[omeo]

Looks like we are jumping the gun:

The linked thread clearly states that the bugs occur in the first game test, the second CPU test, and the sound test. If you look at the [H]'s comparison between the old and new drivers, you will clearly see that the first game FPS is nearly unchanged (101.2 => 103.3) and the second CPU fps is unchanged (12.5), not that that (or the sound test) affects the final score. The massive increases occur in the second, third, and fourth game tests.

Save your conspiracy theories for Deus Ex 2. They'll feel right at home there.

[TimisoaraKill]

Originally posted by s1203372
Is it possible that the fire/explosions use DX9 effects which the GF4's can't create? Do older driver revisions show these effects? Also, does anyone have the same scene so we can evaluate apples vs. apples? Even better, does anyone have an 8500 vs. GF3/4 shot of the same scene?

It will be very interesting to see if Nvidia pulled a quack.

This is only with those 'optimised' 3Dmark driver , the explosions work well with 41.09 and onlder driver even with a GF2 ultra , if you read in 3Dmark forum you can see this by yourself .

[s1203372]

Originally posted by TimisoaraKill


This is only with those 'optimised' 3Dmark driver , the explosions work well with 41.09 and onlder driver even with a GF2 ultra , if you read in 3Dmark forum you can see this by yourself .

Yep, tis true. However, from reading the thread you linked, it quickly becomes obvious that the missing explosions, etc, are not in the games that saw large score increases.

The linked thread clearly states that the bugs occur in the first game test, the second CPU test, and the sound test. If you look at the [H]'s comparison between the old and new drivers, you will clearly see that the first game FPS is nearly unchanged (101.2 => 103.3) and the second CPU fps is unchanged (12.5), not that that (or the sound test) affects the final score. The massive increases occur in the second, third, and fourth game tests.

So, while the drivers DO have some issues with rendering, it's not a quake/quack scenario.

[TimisoaraKill]

Originally posted by s1203372



So, while the drivers DO have some issues with rendering, it's not a quake/quack scenario.

This is what they discovered now, we might see other missing stuff , the bench is big , 2-3 fps from here , 2-3 from another place and this is how you increase a score
This explosions stuff is big and everybody run 3Dmark and nobody see the thing until this guy discovered today , could be how ever a driver issue .
I think that none of the drivers compeared by {H , old and new} and posted here show the explosions (42.63 /42.67 ) so that FPS coment is kinda useless.
The only driver who show all the stuff is 41.09 .

[Chuck232]

I had this too. In GT1 there was no fire, gunfire or anything, but in the CPU test 1, there was. I think they took the screenies from 2 different tests.

[Weedkiller420]

Maybe if yous stop carrying about 3dmarks there wouldent be things like this.

Pls save your conspirecy therious theories the card is not in stores the drivers are not mature

This is not another ati q3 thing.

[Captain Iglo]

Originally posted by Weedkiller420
Maybe if yous stop carrying about 3dmarks there wouldent be things like this.

Pls save your conspirecy therious theories the card is not in stores the drivers are not mature

This is not another ati q3 thing.

the GF4 not available?

in case you haven't noticed yet, this is an issue with drivers for all NV cards, one could even argue that this is even worse than quack, because it's not just an application specific LOD bias (a cheat, sure, but everything is still there, although blurred), but details missing entirely.


any way, the scores obtained with faulty drivers (on purpose or not) cannot be compared to any other scores.

[Paladyr]

Originally posted by Weedkiller420
Maybe if yous stop carrying about 3dmarks there wouldent be things like this.

Pls save your conspirecy therious theories the card is not in stores the drivers are not mature

This is not another ati q3 thing.

Wow, you should really take the time to read through a thread b4 you chime in.

Doesn't look like a quack issue, but I wonder how much of a difference those few FPS would make? I also wonder if maybe there ARE differences in the other tests and no one has noticed them yet.

[Colossus]

I dont believe it, I had tested 3DMarks03 with 41.09 and 42.86 and I saw explosions all the time with fire.

Also look at the frame, they are totally different.

[Paladyr]

Originally posted by Colossus
I dont believe it, I had tested 3DMarks03 with 41.09 and 42.86 and I saw explosions all the time with fire.

Also look at the frame, they are totally different.

Well, perhaps someone with a better reputation than colossus can test this... or perhaps colossus would like to regain some respect by posting screenies using both sets of drivers.