New 3DMark03 Patch 330

[Doomtrooper]

Digit-Life is the english version of Ixbit , you know the site that is using the warez Doom 3 Alpha and making conclusions, that stated the 8500 uses ripmapping, that showed a Ti4600 based cards beating a 9700 in Unreal Tournament 2003...while others showed the opposite.
The site that stated in their forums that Nvidia had the Nv30 in silicon running last April, and that banned members that questioned that very out to lunch comment.

I wouldn't give the site hits..somehow they were able to get double the frames per second on a Ti4600, Crap review site....

Here:

http://www.anandtech.com/video/showdoc.html?i=1683&p=8




Ixbit

vs

Even a Radeon 9700 Overclocked to 350 mhz couldn't beat a Ti4600 according to Uncle Sam and his drivel spreading crew

 

[gokickrocks]

i think its funny how for game test 2 they came up with this conclusion...

The test developers affirm that the NV35 shows snow instead of the dark sky on the beta version of this test. Once again, if users have a perfect sky on their screens on the retail version of the test, then they do not care how some or other programmers make their optimizations. We can see no difference in quality between the versions 3.20 and 3.30 in case of the NV35 and RADEON 9800 PRO. But the NV35 has lost 10-13%, while the R350 has spun up. It looks strange, doesn't it? By the way, have a look at the artifacts when the RADEON 9800 PRO renders the space ship.

...without comparing it to the reference rasterizer (they used the bottom-most spaceship for comparisons)...

http://home.ix.netcom.com/~wicked01/reference840.png

going off topic for a bit...anyone notice how fast the reference rasterizer is on an intel canterwood? it took me less than 5 minutes to render the image

 

[Rookie]

Just heard from a a anonymous source,Nvidia is evaluating the possibility to sue FM ?!

 

[Gubbi]

Interesting, I would think it would be the other way around. FM suing NV over NV's jumping-through-hoops-driver-shenanigans undermine the credibility of FM.

Cheers
Gubbi

 

[Dave Baumann]

Just heard from a a anonymous source,Nvidia is evaluating the possibility to sue FM ?!

The problem for NVIDIA is that at least one of the major OEM's has just sat up and taken note:

http://www.futuremark.com/betaprogram/

Given the timing of Dell's quote, in relation to the last line, it would seem as though they are endorsing FM's recent actions.

 

[Doomtrooper]

Just heard from a a anonymous source,Nvidia is evaluating the possibility to sue FM ?!

 

[andypski]

i think its funny how for game test 2 they came up with this conclusion...
...without comparing it to the reference rasterizer (they used the bottom-most spaceship for comparisons)...

I don't think it's funny at all. Unfortunately I do think it's typical. Somehow I fear for their ability to police image quality in benchmarks when they don't even understand the use of the reference rasterizer...

 

[Himself]

To paraphrase digital life,

"Do you think it's fair that NVIDIA is penalized by 50% based on the image quality?"

WTF? I thought I had heard them all, but that NVIDIA clown car just keeps spitting out more doozies..

 

[RussSchultz]

To paraphrase digital life,

"Do you think it's fair that NVIDIA is penalized by 50% based on the image quality?"

I think, assuming it weren't a benchmark, it would be a relatively fair statement.

Except they ignore the whole clip planes thing--which is where the real increase is coming from.

 

[Joe DeFuria]

I think, assuming it weren't a benchmark, it would be a relatively fair statement.

I disagree completely.

That's a heavily biased statement.

I could be completely biased the other way, and say, "Is nVidia's hardware so screwed up that to get the displayed increases in image quality, it requires a 50% performance hit?"

A fair statement would be something between those two extremes.

Except they ignore the whole clip planes thing--which is where the real increase is coming from.

Agreed, that really is the issue with their "analysis" (for lack of a more applicable term). They apparently didn't read ExtremeTech's article, or FutureMark's PDF....or they didn't understand it. I'm not sure which is worse.

 

[Humus]

Just heard from a a anonymous source,Nvidia is evaluating the possibility to sue FM ?!

Not surprised that they are evaluating the possibility, quite expected given how much it has hurt nVidia's image. But I doubt they will go ahead with it since they have no legal ground to stand on, which I would think their evaluation would conclude too.

 

[stevem]

Not surprised that they are evaluating the possibility, quite expected given how much it has hurt nVidia's image. But I doubt they will go ahead with it since they have no legal ground to stand on, which I would think their evaluation would conclude too.

Their proclivity for litigation need not be based on strong technical merit. There is a group of OEMs baying for blood... This may be a convenient smoke/mirrors gambit...

 

[Rodéric]

Just heard from a a anonymous source,Nvidia is evaluating the possibility to sue FM ?!

Which other 3D firm started releasing average products and suing firms around ?

 

[CorwinB]

Just heard from a a anonymous source,Nvidia is evaluating the possibility to sue FM ?!

I doubt they would do this... They already have enough egg to wipe of their face as it is.

I think they will continue the current insane PR drive, covered by some good "objective" reviewers and by their rabid fanboy base (you know, the kind that used to quote 3DMark2001 to trash the V5).

For what matters, after having used Nvidia cards for a few years (GF2Ultra, GF3, GF4 Ti4600), and being quite happy with them, I sent them an email explaining I would not even consider buying Nvidia hardware again unless I could be sure that they are not cheating in industry-standard benchmarks. And I mean it.

 

[Arun]

Makes no sense at all they'd do that.
Communications doesn't look too good at nVidia, too: About a week after the initial report by Extreme-Tech ( just after the audit of FM, too, I think ) , at least one fairly big head at nVidia was completely left in the dark about the whole issue.

Frankly, maybe someone is seriously considering doing that at nVidia. Expecting it to be approved by the company is ridiculous, however.


Uttar

 

[Gunhead]

Given the timing of Dell's quote, in relation to the last line, it would seem as though they are endorsing FM's recent actions.

I couldn't see a date on it so I don't know how recent Dell's statement is, but at the moment it certainly fits there nicely

Noticed something on the same page that quite needs updating:

How can your company participate?

The BETA Program is targeted for companies that design or manufacture computers or electronic components used in computers. Media and individual participants will not be served through the BETA Program.

So, B3D and ET manufacture or design computers or components, eh, FM website editors?

 

[Himself]

To paraphrase digital life,

"Do you think it's fair that NVIDIA is penalized by 50% based on the image quality?"

I think, assuming it weren't a benchmark, it would be a relatively fair statement.

Except they ignore the whole clip planes thing--which is where the real increase is coming from.

What I would like to know is whatever happened to the video quality zealots that were preaching 32bit over 16bit in the NVIDIA vs 3DFX days.

I think these sites are one glass of lemonade short of worshiping space aliens. Geez.

 

[cellarboy]

Just heard from a a anonymous source,Nvidia is evaluating the possibility to sue FM ?!

Well that's one way to get rid of a dissenting voice. Tie them up in enough expensive litigation for long enough to bankrupt them.

 

[mczak]

Just heard from a a anonymous source,Nvidia is evaluating the possibility to sue FM ?!

Well that's one way to get rid of a dissenting voice. Tie them up in enough expensive litigation for long enough to bankrupt them.

From all we know, Nvidia simply can't win a lawsuit against futuremark (at least if this rumoured-from-an-anonymous-source lawsuit would be based around the "futuremark tried to make us look bad" claim) if the case would actually be ruled (since pretty much _any_ PS2.0 shader is just slower on the NV3x chips than on the R3x0 chips).
Even if the case would never get a ruling, how would trying to put FM out of business (which nvidia might be able to do) help nvidia? They'd lose even the rest of credibility they have left, and I think it wouldn't really help to gain OEM customers. (And by the time FM might go out of business, nvidia probably will have hardware which has competitive PS2.0 performance anyway.)
So, I don't believe they are seriously considering to sue FM. IMHO they could only lose. Best thing they could hope would be that FM would release a new benchmark version which would show the GFFX line in a better light to avoid a lawsuit, but with the OEMs backing futuremark (at least that dell statement looks like that) I don't think this would happen.

 

[mczak]

Looks I was wrong, it's now official, nvidia did not cheat
http://hardocp.com/article.html?art=NDgx
(sorry for that hardocp link, but I couldn't find that press statement at futuremarks website).

Futuremark now has a deeper understanding of the situation and NVIDIA's optimization strategy. In the light of this, Futuremark now states that NVIDIA's driver design is an application specific optimization and not a cheat .

How can you call static clip planes optimization? Answer: you can do that easily if someone holds a gun to your head.
I could accept exchanging shaders as optimizations (if they deliver the same output which they don't in this case), but this just stinks!