nVidia vs Futuremark Continued - Guess what nVidia's doing!

[micron]

This is great news
Let's get everyone back out onto the field the way it should be.

 

[Vince]

What strange comments... It was nVidia, as I recall, last December who quit the 3DMark program and started publicly disparaging the benchmark. T'was nVidia doing the bitching initially, as I recall...

And nVidia has the right (a concept you don't seem to comprehend) to voice their opinion of the company and their benchmark via capitalistic decisions. Thus, as anyone with a hint of education in economics should realize - you lend your support or disaproval with something by investing your money (eg. the investors money) in the products or services that further your means.

It was nVidia's right to pull from FutureMarks, and it's their right (that Futuremarks volentarily extended) to rejoin if it suits their ends.

Your need to attempt to differnetiate between your emotional attachment (or lack there of) towards a particular IHV and attempt see this threw a non-biased view of a comapny in a competitve free-market economy.

T'was also nVidia doing the cheating, which nVidia's never owned up to...yet, although it's a foregone conclusion that they've grossly cheated the benchmark.

You know it, I know it... what do you want? Seriously? Some hint of justification that doesn't exist and a decisive statement from Jen-Hsun Huang that will allow you to post it to a Beyond3D and say "Ahhaaa.. I was right?"

All IHV's cheat, anyone who accepts that a party doesn't in some way is kidding themselves.

Your point about "input from the entire, broad, spectrum of the 3D field" seems inept to me and completely incongruent with any facts I've seen. When was nVidia excluded from providing such input? FM didn't push nVidia out--nVidia left of its own accord.

You seem to totally be missing what I stated. My responce was directed towards those whose comments concerning nVidia rejoining the Beta Group was negative and made comments such as:

Their benchmark can no longer be viewed as an objective performance test

See, this is nVidia's right as the predominent PC IHV to join the FutureMark group if they so choose. Futuremark's underlying principle must be to accuratly gauge the preformance of 3D acceleration on the PC - a key towards doing this is to be impartial towards any IHV and have the cooperation of all who so choose to cooperate.

http://www.futuremark.com/betaprogram/[/url]]The BETA Program is Futuremark's way of co-operating with leading computer hardware and technology manufacturers in order to design unbiased high quality benchmark software. Futuremark provides it's BETA Program Members with specifications and pre-production versions of benchmarks and, in turn, evaluates the observations and remarks provided by the BETA Program Members. This exchange is the foundation of the BETA Program.

As you can see, it is completely integral to Futuremark's success, integrity and corperate integrity to offer a position to all IHVs. It's the IHVs choice if they so desire to be part of this endevour. This is very clear, this is very simple. FutureMarks needs trans-industry support for integrity, the IHV's don't need FutureMarks if it hurts their buisness outlook.

What people are concerned with is the specter of nVidia usurping the benchmark in the future, and that to nVidia "input" means "control." I see no reason that's an illegitimate concern, considering nVidia's behavior to date.

But to go from utter and complete praise of Futuremarks (as well deserved) for their unbiased and unpolitical actions towards exposing nVidia's cheating - to then question their very integrity that was just held in such high esteem based on them [futuremarks] excersizing a fundimental tenet as allowing all IHVs into the program if conditions are met... stinks.

IMO, until nVidia acknowledges its actions and its responsibility, and pledges to chart a future course in a different direction, nothing has been resolved.

Admit it, you just want to hear a comment from nVidia, no matter how shallow, so you can attempt to gain some moral high-ground that just doesn't exist to further your emotional attachment and crusade towards specific companies. It must be, because otherwise you're utterly insane my friend.

Because if you wanted real results, in the real world... you would have come to the conclusion that statements alone mean nothing. Results will come from FutureMarks working with all the major IHVs in implimenting common and real-world routines that are checked by watchdogs like Baumann, Rev, et al. Not blank statements about 'Pledging a new course in the turbulent sea of competition'

 

[Vince]

When futuremark called nvidia on thier bs that was great. When futuremark back down and stoped calling htem cheats that was bad . Nvidia entering the beta program right after haviing its way with futuremark is bad. If nvidia can now controll futuremark with out being in the beta program how much worse is it going to be now that they are back in. Those are my feelings based on prior events.

Bud, you don't know this. As you stated yourself, these are your observations based on what is seen and processed in your (and my) biased minds. The above situation is indicative of nothing and instead can be interpretated in a plurality of ways.

What pisses me off is when people (much like Walt here) allow their passions to override the fact that they don't know the whole situation. To say "nvidia can now controll [sic] futuremarks" is so absurd.

This is a company that allowed Beta members the very privaleges that got nVidia into the situation. This is a company that patched and further exposed the cheating. This is a comapny that has (as Dave said) vowed to make changes and address this issue.

And your going to rip them a new one because of a linguistic difference between "cheating" and "optimizing"? Give me a break... I like my positions to be seemless - if you arbitrarily touch the code, your at fault. nVidia cheats, ATI cheats, they all cheat.

Words and futile conversations over linguistics won't solve this - what will is people like Dave and Rev making sure that FutureMarks does their job. And their job is to provide a cross IHV benchmark, which necessitates that they offer all IHV's (regardless) the possibility to provide input.

 

[Dave Baumann]

What pisses me off is when people (much like Walt here) allow their passions to override the fact that they don't know the whole situation. To say "nvidia can now controll [sic] futuremarks" is so absurd.

Actually Vince, you're doing the same here since you are not aware of the full details. There is still some potential issues, and as a beta member we are going to be following them very closely.

 

[Vince]

Actually Vince, you're doing the same here since you are not aware of the full details. There is still some potential issues, and as a beta member we are going to be following them very closely.

Ahh, I just knew someone would say something like this. Of course I don't know the details, but I'm not the one making charges. I'm just adhering to Futuremark's own statements concerning non-biased inclusion and my own setiments concerning what a company [FutureMarks] in their position should offer to the parties [eg. the ability to join], since it's their benchmark will ultimatly 'make or break' the IHVs and they should atleast have a voice in the process. Now, if it's listened too is a seperate issue.

Hell, in my not so humble opinion - they're all cheaters to some extent. A point that I stated a few times infact. Which is why I specifically stated (and I never had any doubt) that it's people like you who will ultimatly forge the resolution of this issue - for better or worse.

 

[Evildeus]

Actually Vince, you're doing the same here since you are not aware of the full details. There is still some potential issues, and as a beta member we are going to be following them very closely.

Are you? I would dare say no

 

[Hanners]

I would say it's certainly not a bad thing to see nVidia back on the beta program at FutureMark, what probably worries people is that FutureMark have swung very quickly from levelling accusations at nVidia to being best buddies with them.

Now really isn't a good time for FutureMark to start writing demos for nVidia (or any IHV for that matter) if they want to keep their unbiased image. Their reputation has been damaged enough recently as it is, without putting their foot into a whole new can of worms (to mix metaphors) by giving any kind of impression of favouritsm.

 

[Pete]

Actually Vince, you're doing the same here since you are not aware of the full details. There is still some potential issues, and as a beta member we are going to be following them very closely.

Are you? I would dare say no

 

[Dave Baumann]

Are you? I would dare say no

Did I say I did?

However, inevitably I'm aware of a little more than most of you are. As I said before (and at the time of the FM/NVIDIA joint statement) I do have some concerns, and I've spoken to FM's CEO about them, they understand an appreciate those concerns but we both hope they won't be anything more than a concern.

 

[Evildeus]

Saying that "There is still some potential issues, and as a beta member we are going to be following them very closely." seems to say that you are fully aware.

Seems to me that between integrity (and B3D?) and Nv, FM has already made its choice.

 

[Reverend]

I would say it's certainly not a bad thing to see nVidia back on the beta program at FutureMark, what probably worries people is that FutureMark have swung very quickly from levelling accusations at nVidia to being best buddies with them.

I think it is logical to think that a company like Futuremark would want to have all the major IHVs join their beta program. Without the hardware, it would be tough to create benchmarks that are forward looking or implements some crucial features.

I also think that all this does not necessarily mean FM is "buddy" with NVIDIA.

Now really isn't a good time for FutureMark to start writing demos for nVidia (or any IHV for that matter) if they want to keep their unbiased image. Their reputation has been damaged enough recently as it is, without putting their foot into a whole new can of worms (to mix metaphors) by giving any kind of impression of favouritsm.

I'll admit I'm a little uncertain if writing demos for NVIDIA (under whatever arrangements that may exist for this to happen) would be a wise move on FM's part. But I do know a couple of (rather public) FM personnels that are as good as they come wrt integrity.

 

[Dave Baumann]

Saying that "There is still some potential issues, and as a beta member we are going to be following them very closely." seems to say that you are fully aware.

It says that I'm aware of things that could present themselve as potential issues. I've registered those with FM some time ago and they understand them.

 

[martrox]

I have no problem with nVidia rejoining the Beta program. However, I have to say I'm beginning to rethink just exactly what has been done here. Notice that we have heard absolutely NOTHING from FM after the first few days since their capitulation. Now, look at what we have here....think Kyle might just jump back on the 3DMark bandwagon in the near future also?

Originally, everyone thought that the FM rollover was caused by the threat of legal action...... Maybe it's time we rethink that premise...... Because, as has been pointed out many times by many people, the one thing nVidia has a lot of is MONEY......

So, the question might be........which costs more, litigation or Futuremark's soul

 

[Hanners]

I'll admit I'm a little uncertain if writing demos for NVIDIA (under whatever arrangements that may exist for this to happen) would be a wise move on FM's part. But I do know a couple of (rather public) FM personnels that are as good as they come wrt integrity.

Personally, I don't think writing a demo for nVidia shows bias or a lack of integrity on FutureMark's part, but the more important problem for them is the perception of being unbiased in the public eye. That was already damaged by their joint press release with nVidia with regards to cheating, and looking so cosy next to nVidia so soon after that could destroy their credibility (and lets be honest, certain groups are going out of their way to destroy this credibility, and will only try to use this as more ammunition).

 

[Dave Baumann]

Its perfectly normal for FM to write demos - thats part of their revenue stream and they do it elsewhere. However, you've got to question - why would NVIDIA ask FM to do this? Afterall, NVIDIA's demo team is hardly lacking in skill. Clearly its a good way of getting FM to appreciate the architecture of NVIDIA's chips.

 

[Rookie]

From Dave:

why would NVIDIA ask FM to do this?

I agree with dave,it's not reasonable that Nvidia ask FM to write a demo for them?Traditionally Nvidia dont use 3rd party demo.I cant c any reason that Nvidia to use FM's demo.

I remember from where I c a Nvidia man said that they will not return to FM beta program right now.

 

[Evildeus]

Saying that "There is still some potential issues, and as a beta member we are going to be following them very closely." seems to say that you are fully aware.

It says that I'm aware of things that could present themselve as potential issues. I've registered those with FM some time ago and they understand them.

Well, if you say so

So are you giving up intigrity to Nv?

 

[Dave Baumann]

So are you giving up intigrity to Nv?

I fail to understand what you're trying to say.

 

[Arun]

Its perfectly normal for FM to write demos - thats part of their revenue stream and they do it elsewhere. However, you've got to question - why would NVIDIA ask FM to do this? Afterall, NVIDIA's demo team is hardly lacking in skill. Clearly its a good way of getting FM to appreciate the architecture of NVIDIA's chips.

I think you are *very* right there, Dave.
The details about the NV40 microarchitecture *are* rare. There are contradictory reports, and high chances that some things are just wild speculation by people thinking it's "obvious" - not understanding nothing is in this business, eh...
So, on one hand, we've got people thinking things are obvious and they'll say they got inside info. So, well, maybe the obvious is right and they've placed a good bet. Or maybe it isn't, and we'd all look kinda bad if we believed them.

One of the things I'm referring to here, for example, is FP16 in the NV40. Many, and not only speculators on forums, claim it's gone. But who says me their source isn't speculating, or that their source's source is not speculating? Eh...
It does seem to me the register usage problem is one nVidia really doesn't like, and wants to get rid of, so it's likely to be gone - and if it's gone, so is FP16. But whether they can retrieve any advantage of having even a slight register usage hit, maybe a big transistor save, who knows, we don't know either.


Now, in relation to what you posted - even though many things are uncertain, we DO know for sure the NV40 and the R420, even though they are both VS/PS 3.0. , got some fundamental differences.

The first, biggest one is that the R420 is focused on very high speed VS/PS 2.0. - they believe PS3.0. features won't be used for a while. So what they're doing is use the R3xx architecture, add instruction slots & registers, maybe a few instructions or macros to be 3.0. compliant, and for the new features, well... Let's simply say I doubt they'll beat nVidia in dynamic branching, for example.

But does it matter if they are slow in stuff like dynamic branching? Not really. Nobody expects it to be there in games for quite a while.

Now, if nVidia managed to make FM's programmers like the idea of dynamic branching... Eh. Dynamic branching obviously couldn't be used a lot anyway, but still, that's the kind of "feature win" they can get. Same thing for FP16, if it's the case.

I doubt they'd manage to convince FM to use the PPP for anything more than the demo though ( if they even manage that ) - just not enough of a standard, and the IQ differences might be too big to their liking...


Uttar

 

[WaltC]

And nVidia has the right (a concept you don't seem to comprehend) to voice their opinion of the company and their benchmark via capitalistic decisions. Thus, as anyone with a hint of education in economics should realize - you lend your support or disaproval with something by investing your money (eg. the investors money) in the products or services that further your means.

It was nVidia's right to pull from FutureMarks, and it's their right (that Futuremarks volentarily extended) to rejoin if it suits their ends.

Your need to attempt to differnetiate between your emotional attachment (or lack there of) towards a particular IHV and attempt see this threw a non-biased view of a comapny in a competitve free-market economy.

I have no idea what you are talking about...and I don't think you do, either... Of course, nVidia has the right to quit whatever program it chooses to quit--what I was reminding you of is the fact that contrary to your earlier statements it wasn't people, or FM, who started the "bitching" you talked about--it was nVidia.... nVidia is at the root of this problem--not FM, or the people who have observed these events.

Your economic arguments are really nil, you know. Should we let Enron or WorldCom--or for that matter Martha Stewart *chuckle*--off the hook because they are "economic" entities???? Sorry, but being an economic entity in a free market is not a whitewash for cheating and deliberate misrepresentation.

You know it, I know it... what do you want? Seriously? Some hint of justification that doesn't exist and a decisive statement from Jen-Hsun Huang that will allow you to post it to a Beyond3D and say "Ahhaaa.. I was right?"

Good grief--how thick-headed can you be?.... What do I want? A statement by nVidia admitting to *what everybody already knows*, along with a pledge to stop misrepresenting its product performance through the wilful manipulation of benchmarks, would set the record straight and mean that we could all start putting this saga *behind* us. I don't have to see the admission by nVidia to know I'm right--I already do. The fact that they haven't admitted what is so transparent is entirely what's wrong with the picture.

All IHV's cheat, anyone who accepts that a party doesn't in some way is kidding themselves.

Which in no way justifies what nVidia's doing--not even an angstrom. Anyone who thinks it does is entertaining a fantasy. The idea here, you see, is for everybody, *including nVidia*, to stop misrepresenting benchmarks [which of course avoids the fact that as regards 3DMk 03 *nobody* cheated it like nVidia has done--nVidia is quite unique in that respect.]

See, this is nVidia's right as the predominent PC IHV to join the FutureMark group if they so choose. Futuremark's underlying principle must be to accuratly gauge the preformance of 3D acceleration on the PC - a key towards doing this is to be impartial towards any IHV and have the cooperation of all who so choose to cooperate.

Which completely glosses over the point that this was entirely the condition nVidia enjoyed when it chose to be a member of the FM beta team, before it chose to quit, chose to criticize the benchmark, chose to threaten FM legally, and chose to cheat its benchmark software. Nobody's disputing nVidia's "right" to behave in such a disgusting fashion--nobody. But you can't realistically expect too many people to *approve* of that behavior at the same time.

As you can see, it is completely integral to Futuremark's success, integrity and corperate integrity to offer a position to all IHVs. It's the IHVs choice if they so desire to be part of this endevour. This is very clear, this is very simple. FutureMarks needs trans-industry support for integrity, the IHV's don't need FutureMarks if it hurts their buisness outlook.

I still don't understand your point--first you say it's nVidia's right to quit the program--which it is--then you say it's nVidia's right to join the program and pay their dues--which it is--a fact which I don't recall has ever been disputed. You seem to be confusing nVidia's rights to behave autonomously with some sort of failure to understand that FM has all the same "rights" as nVidia--which they exercised when they released their audit report detailing nVidia's cheating (and which, btw, they have not retracted, except for their unfortunate compromise allowing nVidia to mischaracterize its cheats as optimizations.)

If you think that FM is constrained in any way to turn over their benchmark software to nVidia--or any other IHV--you would be very much mistaken. FM has as much autonomy for itself as does nVidia. But autonomy is not the issue here--the issue is what these companies *do* with that autonomy.

But to go from utter and complete praise of Futuremarks (as well deserved) for their unbiased and unpolitical actions towards exposing nVidia's cheating - to then question their very integrity that was just held in such high esteem based on them [futuremarks] excersizing a fundimental tenet as allowing all IHVs into the program if conditions are met... stinks.

Then apparently you have somehow missed a sizable portion of this debate. People became really annoyed with FM when the company caved in and allowed nVidia to attempt (however unsuccessfully) to moderate its behavior not by changing its own behavior--but by moderating FM's initial proof of "cheats" into mere "optimizations." I know I wrote several diatribes berating FM for their lack of backbone at caving in to nVidia's legal bluff so quickly (long before Uttar started this thread.)

If I was FM I would never have sold the integrity of my software so quickly and easily. If necessary, I would have carried the fight to the Internet and sought as much publicity for it as I could garner. If I needed funds to wage a legal fight with those SOB's I would have solicited funds from my program partners and from the Internet at large. nVidia would have quickly seen it had a tiger by the tail instead of mouse--and would have backed off quickly, in my view. They could not have afforded the hornet's nest of negative publicity this thing would have entailed in such a forum. But FM buckled completely at the first hint of opposition from nVidia. They made it seem as if corporations outside of their paying membership had more "pull" than their paying partners. Amazingly, no one in the company seems to have realized what a huge blow this cowardice has hammered into the credibility of the company's software.

So, I don't know where you got the idea of "complete and utter praise" of FM. People were already peeved with the company--and the possibility that nVidia might be reversing itself as to rejoining the program, without ever admitting to having cheated--indeed, all indications are nVidia is still cheating--simply adds fuel to the fire that had already been ignited by FM's weak-kneed capitulation. Hell, if they aren't going to stand behind their software any better than they have done to date--they might as well sell the company to nVidia--because it may be the same difference in the end. This is what people are concerned about. I *hope* my own fears in this regard prove unjustified.