More Nvidia cheating ? 3Dmark2001 now!

[Blastman]

Well, the rabbit hole is getting deeper almost every day. It looks like Nvidia has been dipping in the cheat pot with 3Dmark2001 also -- to the tune of about 11%. See ? beyond3D ?

Basically the GF4 4600 scores dropped from 12,257 to 11,014, not near as much as the 24% in 3Dmark2003 but I guess NV was only getting started in the cheating business with 3DMK2001. If we look at THG VGA charts and move the 4600 down (11.3%), the 4600 falls well behind the 9500pro even with no AA/AF. I?ve adjusted the NV scores accordingly ? 3Dmark2001 ?

9500pro ? ? 12,649
4600 ? ? ? 12,184 ? ? (old ?13,559)
4200 ? ? ? 10,562 ? ? (old ? 11,754)
8500 ? ? ? 10,389

The 8500 is virtually as fast as the 4200 with the adjusted scores.

And there may be more news to follow. It?s reported that the Detonators have a 73 element table of recognized applications and about 50 shaders.

 

[boyRacer]

its called optimizations...

 

[Schadenfroh]

they both cheat, what is the big deal?

btw, dont alot of games have optimizations on ati/nvidia cards?

btw i yet again summon M4H to eat your face!

 

[Yozza]

This whole situation is really said . The fact that they've been doing this long ago with 3D Mark 2k1 as well... makes me sick.

I'm really disappointed in NVidia. I really am. I thought better of them.

The more we scrutinize their drivers, the deeper this hole goes, it seems. I just hope they are brought what they deserve for the deliberate misrepresentation of their products.

Makes me all the happier that I recently dumped my GF4 Ti4600 for a Radeon 9800 Pro, and I would have severe reservations in buying another NVidia product again. The fact that they knowingly and deliberately mislead consumers with these petty "rendering anomalies" is disgusting and heinous.

 

[Viper96720]

Looks like optimizations this time. Did you notice any change in IQ?

 

[yhelothar]

3dmark needs to learn how to produce an accurate benchmark that at least "somewhat" has a repretation of games that are out...

which they never do..

good job nvidia.. make 3dmark die!!

 

[nick1985]

we need M4H...STAT!!!

 

[Megatomic]

Originally posted by: Yozza
This whole situation is really said . The fact that they've been doing this long ago with 3D Mark 2k1 as well... makes me sick.

I'm really disappointed in NVidia. I really am. I thought better of them.

The more we scrutinize their drivers, the deeper this hole goes, it seems. I just hope they are brought what they deserve for the deliberate misrepresentation of their products.

Makes me all the happier that I recently dumped my GF4 Ti4600 for a Radeon 9800 Pro, and I would have severe reservations in buying another NVidia product again. The fact that they knowingly and deliberately mislead consumers with these petty "rendering anomalies" is disgusting and heinous.

I usually save this kind of rancor for child molestors and rapists. Don't you think you are overdramatizing a bit?

 

[squidman]

Yall, dont be hatin on blastman. He done nothing wrong. he wanted to share, is all. THat being said, i would agree with megatomic: you do dramatize. Almost sounds like nVidia snatched your breakfast (makes me mad ). Cmon - let it go. Its marketing. Does intel care that fools spent tons of $$$ on like top of the line PIII 933, which went down once the 1 gig pIII came out? Do i care that i spent 300 bux on geforce 2 gts and 350 (!!!) on 1 stick of 256mb pc 133 ram in july 2000? Naw! If i did, i'd go along saying "taiwan earthquake cheated us all!". Just bad timing, is all.
So cmon, blastman, relax. As long as 3dmark remains the benchmarkers' tool, drivers would be optimized for it. I know my Cat 3.4 is too. Eg: i get 14200 with no aa or af. But with aa and af turned on, no matter how many samples, i would always get 11000+!!! Thats a major cheat right here! But f** it, still a coo card.

 

[Vonkhan]

how deep does the rabbit hole go

 

[peter7921]

Ever since nVidia acquired 3DFX it has turned for the worst. Products are becoming more and more 3dFX like(big huge cards that are over-hyped and under-performing) and ethics are becoming more and more 3DFX like. It looks like 3DFX conqured nVidia from within!!!

 

[nippyjun]

Originally posted by: peter7921
Ever since nVidia acquired 3DFX it has turned for the worst. Products are becoming more and more 3dFX like(big huge cards that are over-hyped and under-performing) and ethics are becoming more and more 3DFX like. It looks like 3DFX conqured nVidia from within!!!

Have a little respect for 3dfx. After all they revolutionized gaming on the pc.

 

[Pete]

"Ethics becoming more and more 3DFX like"--huh?

 

[Corporate Thug]

Originally posted by: virtualgames0
3dmark needs to learn how to produce an accurate benchmark that at least "somewhat" has a repretation of games that are out...

which they never do..

good job nvidia.. make 3dmark die!!

you're blaming the wrong person, it's like blaming a professor for have bad vision when a student cheats

 

[Schadenfroh]

Originally posted by: peter7921
It looks like 3DFX conqured nVidia from within!!!

Victory in death!!!!!

Schaden<--- the lone 3dfxer in a huge world of fanATIcs/nVidiots

 

[yhelothar]

Originally posted by: shady06

Originally posted by: virtualgames0
3dmark needs to learn how to produce an accurate benchmark that at least "somewhat" has a repretation of games that are out...

which they never do..

good job nvidia.. make 3dmark die!!

you're blaming the wrong person, it's like blaming a professor for have bad vision when a student cheats

nahh it's more like...
someone gives you an unfair test to show your intelligence to be able to get a job.. the test is unfair/biased, or has nothing to do with your intelligence.. so you cheated, to make it more fair..

 

[JeremiahTheGreat]

... i do hope that was sarcastic.

 

[nRollo]

I usually save this kind of rancor for child molestors and rapists. Don't you think you are overdramatizing a bit? [

No, he did just buy a 9800, so he's honor bound to defend ATI to the death....


Oh horrors. Can the world survive this video card cheating?!

 

[Dustswirl]

Oh dear!
Anyone used rage3d's built in optimisations for 3dmarks?
i have an nvidia card on another rig and their optimization/cheats call it whatever really feels the same...

(Beware of fake weakness...nvidia is up to something...all this is just $ based (nvidia vs futurmark)
Real benchmarks come from ppl like you and me playing REAL games and it's ppl like us who are the real benchmarkers everytime we jiggle that mouse around!!!)

 

[1ManArmY]

Who cares! Just play the damn games.

 

[Ketchup]

O wait, 3dmark isn't a game? 3dmark is very easy to make optizations for, because it isn't a game, eveything that happens in 3dmark is predictible. ATI does it, nVidia does it. It is not cheating, it's making optimizations. Think about this, you say that "adjusted" the 8500 is as fast as a ti4200. Can you think of any game where the 8500 is even close to the speed of a ti4200? No? Then these adjusted scores are invalid in my book.

 

[Jeff7181]

Oh my god, would you shut up with the cheating stuff already? Gawd... if you want benchmarks, look at the hundreds of benchmarks done in REAL GAMES. There are more than enough here on AnandTech... God forbid you leave the forums and venture into the other areas of this web site.

 

[LesPaul]

If my games look good and run fast, I could care less what 3dmark says about them (good or bad).. I only use 3dmark when I'm stress testing my machine in a gfx environment, or trying to get a feel for how big of a performance boost i get from different tweaks and overclocks.

When trying to decide what video card to buy next.. I don't read all the reviews and go straight to the 3dmark section.. that's usually the section I always skip over, MAYBE give it a quick glance. I wait until people on forums like these, and my RL friends buy stuff first, and listen to what they have to say about it. Reviews and benchmarks do have an influence on my decisions, but not nearly as much as the end users like us, who share thier experiences.

If Nvidia is really cheating on all these things, and not just optimizing.. well, they're gay, but I don't care... I'm still gonna buy thier cards, because they meet my 3 main requirements... They make my games look good, they're fast, and the drivers are stable - no noticable glitches.

 

[nick1985]

Mercenary....Where are you!!!!....

 

[Yozza]

Originally posted by: Rollo
No, he did just buy a 9800, so he's honor bound to defend ATI to the death....

And which video card did I have before the 9800 Pro? A GF4. Before that? A GF2U. I've owned far more nVidia cards than ATI cards, simply based on my criteria for purchase - to buy the better performing product. ATI satisfied that criteria with the R9800 Pro, do you not think?

Yet, you appear to think that I, since now in the possession of an ATI card, am an ATI fanboy? What about when I had the GF4? Would you make the same immediate assumption that I was a NVidia fanboy?

Owning an ATI card right now, has nothing to do with my impressions of NVidia. If anything, I would most likely be even more disappointed and angered if I was using an NVidia card, given that the findings would have had direct relevance to my card. I would be furious with ATI if they had also done what has been uncovered wrt NVidia.

Originally posted by: Megatomic
usually save this kind of rancor for child molestors and rapists. Don't you think you are overdramatizing a bit?

No, I don't, given the countless questionable "optimizations" that have been uncovered.

It's not just this 3D Mark 2k1 incident that angers me; it's the whole situation that has unfolded with nVidia. That's what disappoints me, and perhaps that wasn't made clear enough in my initial post.

They have made an outright denial of any wrong-doing, and continued cheating to cover up for technical deficiencies in their products, which has now been revealed to date back to 3D Mark 2k1 when they were strongly in support of the benchmark.

Sure, it's only a benchmark, and many benchmarks do not accurately reflect game performance, but that doesn't take away from what nVidia deliberately did, and the fact that many do value benchmarks precisely because they offer direct comparison between cards with a given workload, however realistic/unrealistic it may be. Changing this workload unfairly puts the other party at a disadvantage and invalidates the comparison.