3DMark2003 Rendering issues

LMAO!

Mize

 

B3D

It's not an Nvidia forum and i'm not a shareholder. There're plenty thread on this cheat/bug so why not?

 

I agree this driver benchmark hacking is a Bad, Bad Thing to do, but aren't you forgetting something?

Detonators are solid and stabile drivers. I'm yet to experience the first lockup caused by 'Dets and everything just seems to work by the book on the API programming level -- both ogl and d3d like. Sadly I can't say the same about some other vendors' driver sets. ATI has also quite a solid set of drivers these days, but there are still some issues. To me, this seems to be the thing that matters.

Obviously it doesn't make it right nor justify NVIDIA's deeds, but the bottom line is that I'm still happy with my drivers. Sacrifing image quality to the speed god is also a Bad Bad Thing, but currently you've always been easily able to tweak it to the max in the terms of quality.

 

I've yet to experience a lock up on my 9700Pro..

 

Who cares?

I mean seriously, who cares? It is a synth bench that means next to nothing and it garners twenty some comments and all this attention??? bWhat up with that? Isn't there anything intersting in the world of 3D graphics to talk about?

 

*chuckle* No question about it!....

 

I think that in trials you must doubt when there are plausible motives to doubt...

Nvidia history makes more suspicious that driver "bug" (as they say)affecting rendering of 3DMark2003. Though it does seem probable that cheating of this magnitude is not involved in Doom3 demo scores I have no trust whatsoever in Nvidia presenting a truthful account of its performance. I think they used any and all tricks they could to boast those FPS no matter if they reflect a useful improvement or not.

 

Humph=Nvidia PR

 

HardOCP beta members too?

Just curious...since HardOCP seems to dislike 3dmark03 so much, and try to discredit people's results by attacking their motives instead of the data...are they beta members too? Do they have a copy of the beta program to actually see this for themselves?

I read their "non-biased" post on the subject. [H] said they did not care since they did not like the benchmark program anyway. AKA, cheating is ok.

I would agree if the benchmark numbers nVidia received were discarded, unused by any marketing group. That is not the case. The numbers are touted as proving the card's worth. Proof based on a lie.

I am actually more upset at [H] for trying to downplay this than at nV for commiting the evil act!

Sorry, had to rant a moment here...[H] will not let anyone register on their forum...wonder why...

CybrSage

 

I consider myself to be what some people here have refered to as "Joe six pac", (actually more like twelve pac & half a bottle of Brandy) so my question is, how the bloody hell am I supposed to judge the performance of different cards now that there is a question mark over ALL commonly used benchmarks? (imo)

I have relied on reading various reviews, trying to judge from benchmark figures, screenshots & user comments in the past.

Decided to order a Sapphire 9700 non-pro btw.

 

Re: Who cares?

Well in this benchmark we are able to see what this driver is really do for performance gains. We can't see what it is doing in quake, serious sam, or ut2003 because there are no tools to see what is going on.

Benchmarks sell video cards directly to consumers and oems. A card that places well in benchmarks makes big $$$. How do you think Nvidia landed on top.

When you have benchmarks like these people will go out and buy a 5200 or 5600 only to discover the benchmarks run great, but the games are slideshows or they have to tone down the features.

It also makes Nvidia look like the "gold standard" driver leader. How many times have your heard 5800 reviews saying that Nvidia will come out with drivers that give better performance.

Now we know where that performance comes from and we know that it doesn't reflect how well games will perform.

 

Try starting a game of Counter-Strike and press ESC... that should do it if you use Catalyst 3.2. My screen went black and a reboot was required every single time. Other than that, not a single driver related problem 8)

On topic: I think it pretty much looks like a driver bug, not a deliberate "cheat".

 

aha but having been told about when the Cat 3.2's came out, I always use console and type quit thus avoiding it

 

Whether Nvidia are honest may still - for some - be an open issue, but they sure are lucky: other people's bugs screw up graphics, theirs make them go quicker


Derek

 

The cat 3.4's have been released and fix this issue.

Stu

 

I'm no the expert here, but from what I gathered and unterstand, these, let's call them "modifications", go far beyond being a "bug", believe me.

They also easily surpass the (again ) much mentioned quack issue (which was a simple "optimisation" of LOD and/or mipmaps AFIK). Especially in terms of manpower it must have taken to develop and implement the changed planes.

As much as I hate benchmarks, they're (still?) powerfull marketing and sales tools, so in my opinion, what we've got here is outright fraud.

Taking such a risk shows deep desperation and a not-so-healthy attitude towards (potential) customers, reviewers, futuremark (who helped A LOT to make Nvidia what they are tobay) and intelligent human beings.

As for the [H]: well, it's kyle, 'nuff said.

Cheers,
Mac

 

It can not be a bug, it is simply not possible to have a bug where color clears are skipped only when it's safe to and clipping the sky is only done when it's safe to do so on predefined camera path and shows up as soon as you leave it.

 

Oops, I tripped and replaced the engine in my car, damn bannana peels.

 

This is why its soo bad. As now poor "Joe" is given wrong information. About the only way "Joe" can protect himself is to read up on the subject. Never trust one review...look at a few and find the common trends thats usally a safer bet. But we know that "Joe" wont have the time or wont want to spend the time to do that so "Joe" will ask his "PC Guru" or worse ask some one at the store when "Joe" is ready to buy

 

I TOTALLY AGREE!!!!

How many cheeseballs maneuvers does a company have to string together before their reputation suffers? I ain't bound by any kind of moral/NDA/company affiliations so I don't mind saying so; but if it looks like a blatant cheat, and it SMELLS like a blatant cheat, then the odds are pretty good that it's a BLATANT CHEAT!!!! (Sorry, I'm building up me angst again. )

I agree with ya Walt, and I'm larfing me arse off at Mr. DeFuria's comment which I also cynically agree with.

My question now is how long do ya give nVidia to respond? A couple of days, until E3 is over and they have some free time, or do you just wait a few months until the dust settles and they have some new drivers and finally fess up to their earlier debacle? (It seems there pattern of late)

I don't think they deserve this kind of charity, I think they should be held accountable and I'd feel a WHOLE lot better if'n they could give SOME kind of response by now...it's been a few days and this just seems like a big deal to me that they're trying to wait out.