Help me understand Kyle

[WaltC]

You forgot "dark and evil". It should be "Perez exists at the dark and evil heart of the nVidia PR machine".

Heh-Heh... I prefer to simply think of him as "smoking something hallucinogenic"...

 

[digitalwanderer]

You forgot "dark and evil". It should be "Perez exists at the dark and evil heart of the nVidia PR machine".

Heh-Heh... I prefer to simply think of him as "smoking something hallucinogenic"...

No, no...that's me. (I haven't had any cigarettes or cigars though! )

 

[Sxotty]

Walt
What I was saying is this, Nvidia is working with devs to optimize so devs will not try to intentionally break things.

Nvidia and Valve are working together NOW, this means the shaders for the FX series will be in the game to begin with and won't need to be replaced in the benchmark by the drivers. You see what I mean, if nvidia wants to replace them in other games and the dev is ok, then nvidia will just give them to them and then they will be included. If they do not negatively impact IQ, and nvidia did them instead of having other people do them they would be grateful. Think valve spending 6x more effort on FX (or whatever they said), vs. Nvidia spending the effort and saying "here ya go valve, hot shaders fresh from the dark heart of..." you get the idea .

Yes the compliler exists as has been beat to death over in the 3d technology and hardware area.

Yes the PR department tried to say somehow that

These techniques are applied with a fingerprinting mechanism which evaluates shaders and, in some cases substitutes hand tuned shaders, but increasingly generates optimal code in real-time.

Which basically means, when we can't get the desired results automatically (generically) then we substitute hand tuned shaders (cheat)

 

[Joe DeFuria]

You see what I mean, if nvidia wants to replace them in other games and the dev is ok, then nvidia will just give them to them and then they will be included.

But this is not sustainable.

Are devs going to have to worry about getting their shaders "nvidia approved / optimised" every time the come up with a new shader, or modify an existing one?

Will nVidia offer "this service" to the hudreds of developers and Lord knows how many shaders? Or only the "popular" games and of course, those used for benchmarks?

If theyWhich basically means, when we can't get the desired results automatically (generically) then we substitute hand tuned shaders (cheat)

Correct. And it is perfectly legitimate for FM to refuse hand tuned shaders. (I'm not sure if you disagree with that or not. )

 

[Hanners]

Correct. And it is perfectly legitimate for FM to refuse hand tuned shaders. (I'm not sure if you disagree with that or not. )

At the end of the day, I would say it's legitimate for any application/game/benchmark developer to do what they like, as long as they remember that they have to live or die by the choices they make.

It's perfectly acceptable for the developer of any title to refuse hand-tuned shaders - At the end of the day it is their work and they can do what they choose with it. Obviously, in the real-world most developers (of games at least) would welcome them with open arms, but I don't like nVidia's attitude that they should be allowed to force anyone who doesn't agree with them to take their route and play by their rules. They are there to run the software as the developer intended, not rewrite the software for them.

 

[nelg]

I have been re-hashing the issue with Kyle and someone from AVAULT.COM in this thread, if anyone wants to join the fun ..........................
http://www.hardforum.com/showthread.php?s=&threadid=688550&perpage=15&pagenumber=1

Those forums are currently undergoing a minor database implosion, it seems

I think that the forum is down to purge John Reynolds post and ban him. Kyle and Brent just cannot seem to be able to counter his reasoning.

 

[WaltC]

Walt
What I was saying is this, Nvidia is working with devs to optimize so devs will not try to intentionally break things.

Which devs try to intentionally break things, though?... I can see nVidia "working" with devs so that the devs won't unintentionally break things in nVidia's drivers, because the devs are writing code congruent to particular API specifications which the devs have wrongly assumed nVidia's drivers correctly support. I can see nVidia working with devs to rectify such misunderstandings, certainly... But I really can't see devs deliberately trying to give nVidia's drivers a hard time specifically, since plain old bad API code would be just as damaging to ATi's drivers at the same time.

Which basically means, when we can't get the desired results automatically (generically) then we substitute hand tuned shaders (cheat)

I don't really care if they do that in a 3d game (I do care about other issues like trilinear filtering in 3d games, but that's a different issue.) I leave the questions of what nVidia's drivers do in 3d games to the short list of hardware reviewers whose judgment I trust. What's important here is that one of the legalities nVidia has voluntarily agreed to accept as a member of the FM program is that any and all driver-detection for 3dMk03 is off the table. nVidia knows they are not supposed to be doing it, but they are doing it anyway. Hence, FM has no option but to issue detection-killing patches if it wants to maintain control of its software, not to mention the credibility of its software, because unlike 3d games, 3dmk03 is a 3d benchmark suite often used to compare competing 3d hardware, and actually has no other purpose (even if DP can't understand the difference between 3dmk03 and Q3.)

 

[volt]

I'm VERY curious if NVIDIAs attitude will remain as is when NV40 is introduced. I'd like to point out that making not-so-legal workarounds to boost performance costs lots of money, not to mention all the marketing that goes along with it.

Is it in NVIDIAs future plans for the NV4x to actually keep substituting shaders for their own? What about Non TWIMTBP games? Shaders were developed to ease developers job and create effects they planned on including without IHVs support. I hate to imagine what is going to happen If NVIDIAs "tomorrow" GPU is going to be as weak as the NV30 / NV35 and the drivers will keep illegally optimizing games and benchmarks.

Or maybe NVIDIA spent all that money just for the NV3x series, who knows. IIRC, they were the ones saying 0.13u process and Low-K were bad

 

[Bouncing Zabaglione Bros.]

I'm VERY curious if NVIDIAs attitude will remain as is when NV40 is introduced. I'd like to point out that making not-so-legal workarounds to boost performance costs lots of money, not to mention all the marketing that goes along with it.

Depends whether R420 continues to spank NV40 the same way R3x0 spanks NV3x.

Nvidia will stop cheating and lying when they can get back to claiming to be "the fastest" at everything. Frames per second appear to be the only string to the Nvidia bow, which is why they are cheating so hard to try and get it back.

Looking at the way Nvidia have historically developed their products and caried previous design decisions onto newer parts, while relying on smaller chip processes and faster RAM to boost their performance, I suspect the spanking by ATI will continue.

 

[nelg]

Walt
What I was saying is this, Nvidia is working with devs to optimize so devs will not try to intentionally break things.

Are they also working with devs to make sure that it hampers performance on ATI hardware, as rumors have implied?

 

[WaltC]

I'm VERY curious if NVIDIAs attitude will remain as is when NV40 is introduced. I'd like to point out that making not-so-legal workarounds to boost performance costs lots of money, not to mention all the marketing that goes along with it.

I fully agree with you--this is a dead-end, no-win strategy for nVidia. "A rat chasing and eating its own tail," is the kind of mental imagery it invokes for me... It's got to cost a lot of money and time that one would think might be better plowed into R&D on future products. I think the company is running the very high risk of indoctrinating itself into a pattern of behavior that it may not be able to switch off as easily as a light switch, when if ever the time comes. nV4x? If anything, the company's present mode of behavior would lead to me to think it is anything but a radical departure from nV3x.

Is it in NVIDIAs future plans for the NV4x to actually keep substituting shaders for their own? What about Non TWIMTBP games? Shaders were developed to ease developers job and create effects they planned on including without IHVs support. I hate to imagine what is going to happen If NVIDIAs "tomorrow" GPU is going to be as weak as the NV30 / NV35 and the drivers will keep illegally optimizing games and benchmarks.

Or maybe NVIDIA spent all that money just for the NV3x series, who knows. IIRC, they were the ones saying 0.13u process and Low-K were bad

Very good and astute questions. nVidia over the past year if anything has demonstrated it can be as stubborn as a mule, and that it becomes fixated on specific doctrines, approaches, and technological directions quite easily. Of course, this can all be explained if we examine the idea that the company has never been particularly strong in gpu architecture design in the first place, but has traditionally relied on the adoption of 3rd-party FABs, ram makers, etc., to give it a competitive edge in the 3d marketplace since 1999.

Personally, I think the company's been very remiss in some key areas. What, for instance, would have made nVidia think that a reference design for nV3x equipped with "FX-FLOW" might have been desirable from the standpoint of system OEMs like Dell? One of the reasons nVidia overtook 3dfx was because of the unpopularity of the V5.5K reference design: it was too long, too hot, and consumed too much voltage in their boxes when compared to nVidia's nv10/20 reference designs of the time. 3dfx really took a beating in the OEM markets because of that. Yet, with nV30U, nVidia foists a reference design that is hotter, far noisier, almost as long, and consumed more power than even 3dfx's aborted V56K; and to top it off sported a double-wide backplane as well. How was it nVidia could have failed to ascertain the lessons learned by 3dfx regarding the importance of the system OEM markets in the scheme of things? I think nVidia's got some real blind spots, and as such it will certainly be interesting to see what sort of departure from nV3x nV4x will turn out to be.

 

[nelg]

Maybe this will help you understand Kyle AnteP.......

quote:
--------------------------------------------------------------------------------
Originally posted by John Reynolds
Nice pot shot, Kyle. But it usually doesn't take you too long before the insults start flying.

That aside, what if someone coded a nice little synthetic test that stresses stencil performance for modern graphics boards. Wouldn't you think that might be a nice way to evaluate how a 9800 XT might do in Doom 3 compared to a 5950 Ultra? Or is there a published game available now that makes heavy use of stencils (if there is, I'm not aware of it)? Don't you think readers who want the best board for Doom 3 are interested in such tests? I certainly would be!

It's all about giving your readers good information so they can make informed, intelligent purchasing decisions, not putting silly little blinders on and going all anti-synthetics. Sure we buy these boards to play games, but why we buy the particular boards we do is based on expectations of performance (and features) derived largely from reviews, online or print. The more information such reviews give the consumer the better he/she's future purchase will satisfy their expectations.
--------------------------------------------------------------------------------



Oh that was not a pot shot, that is an honest request that leaves you will nothing to say Mr Senior Editor and if you read carefully there was not an insult in there. I guess you see me calling the facts of what you have done into question as an insult? That is funny.

"Kyle is such a meanie! He asked me what I had done with my Senior Editor status at Beyond3D in terms of addressing the situation this year!!! What a bastard."

If it is all about giving readers good information, maybe you ought to start doing that as I have seen nothing from you.

I will stand behind every word we publish and will proudly do so. The work we do is damn good and I am proud of everyone associated with it. Little whiners and fingerpointers like you will not ever make a difference till you figure out that the "real world' does not give a damn about much anything posted on a message board.

Now will you go back to your editor desk and whine there? (Was that too rough and insulting for you? Did I make you cry?)

Can you say egomaniac.

 

[Miksu]

Can you say egomaniac.

I too read that thread and noticed that post. I really don't get it. How can one person, who really don't know how to discuss politely be in a position he is nowadays? Well, he doesn't have any position but he thinks he has one and acts like "I'm king of the world". ...

 

[Doomtrooper]

Well that is easy, today running a web review site has nothing to do with content...the reader base is usually 10-16 years old at [H] so all you really need to do is post some neon cases reviews, throw roxzorzzzz and kickass and some other cool internet lingo into the review and you have a winner.

Kyle is a businessman, don't ever think otherwise.

 

[Sxotty]

Correct. And it is perfectly legitimate for FM to refuse hand tuned shaders. (I'm not sure if you disagree with that or not. )

I agree I completely understand why and support the refusal of hand tuned shaders (sorry I was unclear and never addressed this issue).

The only thing I was saying to walt is that Nvidia is not trying to be crafty while detecting games, so it will not be quite so easy to break.

So if Nv does help write some shaders for some games the devs when they replace them will have a folder with the nvidia specific shaders, and they will make sure they are still good, see what I mean... not that nv wil include them in the drivers that is unsustainable and will get silly quick (although now I am sure they only do a few in games that are often benchmarked)

Thanks Joe for not saying I said something I didn't (that always annoys me when a weird statement is attributed to me).

 

[Sxotty]

Are they also working with devs to make sure that it hampers performance on ATI hardware, as rumors have implied?

Well I think you know the answer look at Valve, Nvidia is sure working with them, but how did ATI's cards compare?

Whether the rumors have any validity I doubt any major game dev would do it.

 

[jjayb]

Are they also working with devs to make sure that it hampers performance on ATI hardware, as rumors have implied?

Well I think you know the answer look at Valve, Nvidia is sure working with them, but how did ATI's cards compare?

Whether the rumors have any validity I doubt any major game dev would do it.

Tiger woods? Don't remember what ever became of that whole situation. Correct me if I'm barking up the wrong tree.

 

[Sxotty]

We are really diverting this thread and I apologize this is the last OT post for me.
I do not think it is in the best interest of a dev to purposely make a game run bad on any system. This is clearly a shortsighted thing to do, and while a small dev, or a subsidary that can be canned might do it, no respectable real dev would do it b/c in the long run it would piss people off and alienate some portion of their customers.

 

[John Reynolds]

Maybe this will help you understand Kyle AnteP.......

quote:
--------------------------------------------------------------------------------
Originally posted by John Reynolds
Nice pot shot, Kyle. But it usually doesn't take you too long before the insults start flying.

That aside, what if someone coded a nice little synthetic test that stresses stencil performance for modern graphics boards. Wouldn't you think that might be a nice way to evaluate how a 9800 XT might do in Doom 3 compared to a 5950 Ultra? Or is there a published game available now that makes heavy use of stencils (if there is, I'm not aware of it)? Don't you think readers who want the best board for Doom 3 are interested in such tests? I certainly would be!

It's all about giving your readers good information so they can make informed, intelligent purchasing decisions, not putting silly little blinders on and going all anti-synthetics. Sure we buy these boards to play games, but why we buy the particular boards we do is based on expectations of performance (and features) derived largely from reviews, online or print. The more information such reviews give the consumer the better he/she's future purchase will satisfy their expectations.
--------------------------------------------------------------------------------



Oh that was not a pot shot, that is an honest request that leaves you will nothing to say Mr Senior Editor and if you read carefully there was not an insult in there. I guess you see me calling the facts of what you have done into question as an insult? That is funny.

"Kyle is such a meanie! He asked me what I had done with my Senior Editor status at Beyond3D in terms of addressing the situation this year!!! What a bastard."

If it is all about giving readers good information, maybe you ought to start doing that as I have seen nothing from you.

I will stand behind every word we publish and will proudly do so. The work we do is damn good and I am proud of everyone associated with it. Little whiners and fingerpointers like you will not ever make a difference till you figure out that the "real world' does not give a damn about much anything posted on a message board.

Now will you go back to your editor desk and whine there? (Was that too rough and insulting for you? Did I make you cry?)

Can you say egomaniac.

Post in forums and hopefully know what you're talking about from time to time or run your own site and show the world your ignorance on an almost daily basis? Hmmm, tough choice.

Egads, I hadn't seen that response since the [H] forums have been down most of the day. And though he's obviously being a wanker, he has a point in that I really don't do much for Beyond3D. Over the past 1.5 years I could've been far more aggressive in pursuing AIB vendors for review boards, but with a full-time job, new house, a newborn baby, and, perhaps most importantly, a waning interest in PC hardware, I just don't want to commit my free time to testing and writing. And, funnily enough, I've removed both B3D and SimHQ from almost all my sigs last week.

Of course Kyle's comments completely dodge whether or not what I write, whether it be in forums or a front page, is factual and/or logical. And I'm certainly not going to waste my time replying to him since his post gives me one big fat "meh" response.

 

[Doomtrooper]

and, perhaps most importantly, a waning interest in PC hardware

I am also in that boat...must be the age factor and twin girls don't help either.