ExtremeTech Article Up

[Doomtrooper]

Using common sense here....

Do you actually think that the 5900 is so advanced that it can warp through time to the future, know what is being rendered, warp back through time and insert clipping planes in areas it knows it can.

Sorta Like 'Back to the Future'..does a FX5900 have a Flux Capacitor on it

 

[Joe DeFuria]

When did the reviews have been made? And when did the information have been given to Nvidia?

Where's the statement from nVidia telling all their web review sites to pull 3D Mark scores?

Don't you say, innocent proven guilty?

No, I say present all the truthful, repeatable, and verifiable evidence and let the trial begin. Not, withhold evidence while scores are being published and drivers are being downloaded.

BTW, you did see that FM has said they don't know if it's cheating or not.

Huh? Right. I said they have no official posistion of if it's cheating or not. This doesn't change the fact that
1) Drivers are doing something they shouldn't
2) These drivers are officially sancioned by nvidia, are available for download, and reviewers are publishing scores based on them.

Yes i have that in mind. But i have also in mind that the information we have shows us that Nvidia didn't have much time to respond,

They DID respond with "probably a driver bug." The did NOT respond by pulling the drivers and requesting scores to be withdrawn.

Futuremark did use some exceptionnal means...

FM did NOTHNING but corroborate ExtremeTech's findings.

and that the article comes out 3 days after the launch of a new product from NV.

??? The article came out shortly after the new drivers were officially distributed. It's not FM's or ET's fault that nVidia paired a new product launch with new drivers.

Futuremark still don't know, but FM accept the publication of the news. Odd at best i say.

FM accepts the publication of the EVIDENCE because the evidence is 100% truthful.

I don't really want to rego in this, but yes there's some issues with NV drivers (it's most probabky cheating),

Again, issues in publically available drivers

no FM don't know if it's a bug or not, no NV/FM didn't have enough time to go further in the investigation,

Regardless, the evidence is true, reproducible, and verifiable.

yes the article is well documented, yes the article has been writen before any conclusion from FM (in a rush?),

Since when does anyone need a "conclusion" from FM to publish anything?

Pardon me, but where the hell were you when Kyle was dissing 3DMark03, based on nVidia's accusations, and not even BOTHERED to ask FM about the validity and relevancy of nVidia's claims?

yes it has been done with special authorisation from FM.

Right, because they cannot publish stuff using their developer tool without permission. What's the problem?

 

[Joe DeFuria]

Sorta Like 'Back to the Future'..does a FX5900 have a Flux Capacitor on it

Dunno...but based on the cooling solutions of the FX series, they probably consume about 1.21 GigaWatts of power...so you never know.

 

[Doomtrooper]

I never thought of it that way

 

[RussSchultz]

Using common sense here....

Do you actually think that the 5900 is so advanced that it can warp through time to the future, know what is being rendered, warp back through time and insert clipping planes in areas it knows it can.

Sorta Like 'Back to the Future'..does a FX5900 have a Flux Capacitor on it

It doesn't need to know the future. All it needs to do is know about the current frame (or, presumably, it could rely on the previous frame). Of course, it would be on the same level as that mythical HSR driver that 3dfx had, but you don't need to break any laws of physics to accomplish what you're suggesting.

(Yes, I know, it would need to buffer an entire frame, but it might be an interesting topic of discussion to guess whether a 3.0 HT capable P4 could come up with reasonable clip planes in real time either by analyzing the previous frame's geometry, or the by buffering the current frames.)

 

[muzz]

They are not going to pull scores, or withdraw the drivers for 1 reason IMO ( and thats ALL it is)....

They are making $ on those scores, as the avg joe blow doesn't even know this exists, the only thing they see is it kickin butt....and THEY make up a big part of the sales...... and lets not forget there is a trickle down effect to lower grade versions of the card.
IMO they will NOT pull it until they HAVE to ( if ever)....

It is pretty obvious that something is wrong, regardless if the accusations are true, only thing is that it is only the online enthusiasts that know about it.

 

[Doomtrooper]

Buffer a entire frame, examine it then put a clipping plane in, all while the benchmark/game is being used..seems too far fetched for todays technology, but I'm not a graphic card designer either.

 

[RussSchultz]

Buffer a entire frame, examine it then put a clipping plane in, all while the benchmark/game is being used..seems too far fetched for todays technology, but I'm not a graphic card designer either.

Many older cards relied on the PC for transform, lighting and clipping (of triangles that ended up off the edge of the viewport). Not that its entirely the same thing, but its similar. I don't think you could do it in the general case, but if you know and the work spent culling saves more work running shaders/filling then its a win. At worst, I think you'd have to have 1 frame of 'lag' a la the ATI MAXX to maintain the pipeline of data going to the card. (the flip() would move 2 frames ago to the front buffer, 1 frames ago is being analyzed and downloaded/rendered to the backbuffer, the current frame now begins collecting data)

Of course, that really doesn't explain the 3dmark3 things get all wonky when you go off the rail, but its still an interesting idea.

 

[Evildeus]

Where's the statement from nVidia telling all their web review sites to pull 3D Mark scores?

Really So each time there's a bug on a application we should see this statement no? I haven't see any till now.

But well, i would be pleased that NV begin.

No, I say present all the truthful, repeatable, and verifiable evidence and let the trial begin. Not, withhold evidence while scores are being published and drivers are being downloaded.

I see, and the evidence is that there's a problem, thet FM and NV doesn't know what it is (or they say so), and that the trial has already been conclude by people, ie GUILTY.

Huh? Right. I said they have no official posistion of if it's cheating or not. This doesn't change the fact that
1) Drivers are doing something they shouldn't
2) These drivers are officially sancioned by nvidia, are available for download, and reviewers are publishing scores based on them.

And that special process in this case, saying giving confidential information before any futher investigation contrary to the rules of FM.

They DID respond with "probably a driver bug." The did NOT respond by pulling the drivers and requesting scores to be withdrawn.

I didn't say they didn't i said they didn't have time to go far. And see above.

FM did NOTHNING but corroborate ExtremeTech's findings.

Really? I did think that it was an FM product, and that ET was a beta member. Thx to you to clear my mind: It's an ET product with FM as a beta member

??? The article came out shortly after the new drivers were officially distributed. It's not FM's or ET's fault that nVidia paired a new product launch with new drivers.

Really, then there's an issue cos FM/ET said they were aware 1 week before and moreover that NO investigation has been made with NV (oh yes "it's probably a bug" is really some investigation) and no investigation has been conclusive at FM right now.

FM accepts the publication of the EVIDENCE because the evidence is 100% truthful.

And in that specfic case it's a special treatment, normally it would have been investigate in confidentiality, then when the conclusion comes out there would have been an official statement.

Regardless, the evidence is true, reproducible, and verifiable.

Yes, but it's the conclusion that is more important, because it's not able to see it on 3Dmark without beeing a Beta member. And the only conclusion i see is: ET saying NV is guilty and FM saying, we don't know look at ET. 8)

Since when does anyone need a "conclusion" from FM to publish anything?

Since it's being established in the beta program, using confidential information. But well it doesn't seems to have been include in your requirements.

Pardon me, but where the hell were you when Kyle was dissing 3DMark03, based on nVidia's accusations, and not even BOTHERED to ask FM about the validity and relevancy of nVidia's claims?

I was not here, so what? And what's the releavance of this? Does that means that what i'm saying is false?

Right, because they cannot publish stuff using their developer tool without permission. What's the problem?

Well, if you don't see it, i can't do much for you.

 

[Unit01]

Someone needs to use some logic and learn some more english and re-read a certain article. The discussion is escalating into the regular 1.repeat 2.ignore 1.repeat 2.ignore(doesn't know what it is about anymore, subject change) discussion where the numbers represent people.

 

[Dave Baumann]

Yes, but it's the conclusion that is more important, because it's not able to see it on 3Dmark without beeing a Beta member. And the only conclusion i see is: ET saying NV is guilty and FM saying, we don't know look at ET.

ED, I'd say you are trying to read too much into what they are saying. I would say they are not saying they don't know what it is, but just not commenting.

As for the 'rules of the Beta membership' you keep quoting, in reality bringing press onto the Beta group is a new thing and the rules are being made up as they go along to a certain extent. We joined because we want to help shape the benchmark, however inevitably we are all now the policing of it to a certain extent - if it helps to keep the use of the benchmark an honest one then that is a good thing, but if thats the case that part of what the Beta press members are to do (as I said, thats not why we joined) then I'm all for the Beta membership being opened to more press (as I have already suggested to Futuremark).

 

[Reverend]

ED, I hate accusing anyone of "cheating" but insofar as the ET article and our "backing up" of it is concerned... are you concerned with what this means? Why are you so much more concerned with "confidentiality" and whatnots than the information ET presented?

 

[Evildeus]

ED, I'd say you are trying to read too much into what they are saying. I would say they are not saying they don't know what it is, but just not commenting.

Well, then i would prefer that they comment because it would end up this endless discussion . You seems to know more (as a beta member ) but the official statement is:

Right now we are gathering more information about the subject and may comment to this later.

PS: What does ED means?

As for the 'rules of the Beta membership' you keep quoting, in reality bringing press onto the Beta group is a new thing and the rules are being made up as they go along to a certain extent. We joined because we want to help shape the benchmark, however inevitably we are all now the policing of it to a certain extent - if it helps to keep the use of the benchmark an honest one then that is a good thing, but if thats the case that part of what the Beta press members are to do (as I said, thats not why we joined) then I'm all for the Beta membership being opened to more press (as I have already suggested to Futuremark).

It's a good thing to have the press as beta members, as you are the first people interested in benchmarks. If the rules have changed, let us know

If you are saying that anytime you see issues with FM benches you will show it to the public whatever is the company involved, then it's new and good information.

But if it depends on the issues and the company, then there's a problem of interest, as i said before.

 

[Seiko]

fighting fire with fire?

Now I know this is going to attract a lot of heat but as Nvidia is allowed to win the benchies due to certain and perhaps unethical optimisations what if ATI added the same optimisations to their drivers? It could be selected in the control panel and of course should be broadcast and documented in full. Although I think this may cause unintentional damage to 3DMarks credibiltity it would show all the sites two can play at that game? At the very least the sites would have to state that the optimisations aren't really going to translate into improved game framerates but hey, at least Nvidia is shown to be using the same optimisations and as they didn't allow the option to be controlled they would certainly loose a lot of credibility! Of course I'm assuming that if ATI added the same style of optimisations to their drivers we'd see a marked improvement in the scores and would probably level the playing field?

Other than that I dont see ATI with any options, at the moment Nvidia are clearly enjoying most people seeing them win the 3dmark benchies and won't care if a few specialised sites throw a little mud because of the optimsiations in place. As I hate to see a company pretty much pull the wool over it's consumer eyes I really do hope this whole affair isn't allowed to rest.

I hope ATI has the creativity to turn Nvidias under handed tactics into Nvidias own worst nightmare!

 

[RussSchultz]

Re: fighting fire with fire?

Now I know this is going to attract a lot of heat

Maybe it wouldn't attract so much heat if you didn't post it in multiple threads?

 

[Seiko]

Re: fighting fire with fire?

Now I know this is going to attract a lot of heat

Maybe it wouldn't attract so much heat if you didn't post it in multiple threads?

Perhaps? But then I wanted to actaully start a new thread but as I think it could be answered or discussed from either a cheating ethical viewpoint or solely Extremes article, hence the dual post.

Anyway, I'm sure you're big enough to get on with it

 

[Evildeus]

ED, I hate accusing anyone of "cheating" but insofar as the ET article and our "backing up" of it is concerned... are you concerned with what this means? Why are you so much more concerned with "confidentiality" and whatnots than the information ET presented?

I have a NV product, even if it's a old GF4 TI i'm concerned.

Can't i have many other concerns? Should i be only single minded?

Why? Because it's confidential information, how i know if in the future, any issues will be handled the same way? How do i know if the current relationship between ET/FM/B2D/ATI/MATROX/etc. and non beta members won't have effects on the "exceptional" providing of issues? How do i know if all non beta and beta members will be covered the same way?

Perhaps that doesn't concern you, but it does concern me. Well yes the issue is important, and my concern is also important to me. Is it forbidden?

 

[Randell]

Why? Because it's confidential information, how i know if in the future, any issues will be handled the same way? How do i know if the current relationship between ET/FM/B2D/ATI/MATROX/etc. and non beta members won't have effects on the "exceptional" providing of issues? How do i know if all non beta and beta members will be covered the same way?

It is not confidential! Why on earth would identifying suspect action in a benchmark be confidential to a non-beta member!?? nVidia are not a member any NDA rules do not apply to them, theri drivers or their products.

Why do you think nVidia should be protected? Who is protecting the OEM's and consumers if discovering these things is hushed up under NDA protection?

And in cases like this, whistleblwong is to be applauded, not anallysed to the 9th degree in orer to find some rule to hang the whistleblower by. What kind of community do you want to live in where you are more concerned with punishing the messenger rather than the 'wrongdoer'?

Can B3D elaborate on how the rules would govern anamolies being discovered in member situations?

 

[Dave Baumann]

Why? Because it's confidential information, how i know if in the future, any issues will be handled the same way? How do i know if the current relationship between ET/FM/B2D/ATI/MATROX/etc. and non beta members won't have effects on the "exceptional" providing of issues? How do i know if all non beta and beta members will be covered the same way?

I'm getting a little tired of this.

First off, as people have said to you time and time again for non-beta members the only people who deem confidentialiy or not are Futuremark.

Do you think that any of the Beta participants are going to do something that can be found within the developer releases given they know media have the same release? Frankly that would be stupid of them to attempt something like this in the first place.

As for handling it, what did we do? We looked at the issues informed NVIDIA that we have seen them, supplied them with screenshots (and broke NDA in the process) and asked them for an explaination. Do you think we would do anything different with anyone else? No, we would hadle it the same, although with a beta member we won't be breaking NDA in the process.

Frankly everyone here has been overly fair in the handling of the situation and I'm getting a little sick and tired of the insinuations that we wouldn't do the same for anyone else.

 

[Evildeus]

Well randell, I didn't do the rules, so don't shoot on the messenger and then why did FM got mad @ B3D if there's no NDA?

Who is protecting the consumer? FM, they are the provider, they are the one protecting their product and therefore the consumer.

Who is protecting me as all the beta members are protected by NDAs?

And if they don't apply it to non-members but only to members, well it's not a society when i would like to live in.

As for messengers, i don't think they are neutral, so i don't see why i should stay neutral to them.