3DLabs Cg Rebuttal

Doomtrooper said:
If you are going to disagree provide more to this debate then replying to me about Nvidia is evil crap, I posted my opinion on this subject along with quotes from the OGL and 3Dlabs (what this thread is about BTW).
If you don't like it, tough luck.

What is it with you and this tough luck stuff? Where did I mention that I didn't like the quotes? When you can't add anything more to a conversation you cross your arms and huff and puff now?

Let me make it clearer for you. I AM part of a dev team (as insignificant as we are though) so I at least think I am trying to get a clear grasp of CG and its implications versus your hunt for the evil that is NVidia. All you have to do is address my points. That's all, it's not rocket science here. ;)

But apparently this discussion is at an end with you because for you, NVidia = the devil. Btw, our game DOES have specific ATI extensions so please check your fanboy comments at the door because obviously you are not really trying to understand CG but rather only find its faults.
 
Sabastian said:
Another good point that Doomtropper brings up is the fact that there are no other graphic chip companies having praise for nvidias Cg. In fact we have an official rebuttal from 3Dlabs.
Why would other graphic chip companies do something altruistic for nvidia?
Why would nvidia speed up development of software so that it can run better on other companies hardware just as well? That would truly be an act of altruism on nvidias behalf.
Because more developers would use it. I dont see why that is altruistic. Software taking advantage of modern hardware is in nVidias best interest.
 
Doomtrooper said:
Since they were the ONLY company wanting to charge licensing fees for extensions, the company that took the Open out of Opengl
...
I have a real big problem charging licensing fees to a API that was never meant to have them :devilish:

#1) S3 originally charged licenses for S3TC
#2) SGI originally charged licenses for hardware vendors to implement OpenGL.

So, unless you're rewriting history, and assuming that SGI has made OpenGL freely available for only altruistic reasons (and NVIDIA couldn't possibly do such a thing), I think you need to get off your high horse and quit preaching your bias--Its really getting tiresome.
 
This C for Graphics (CG) how optimised is the output, compare to hand code of Vertex and Pixel shader ? Well assuming common denominator of nVidia card ?
 
Proprietary extensions for a common function (GL_NV_vertex_program) is not texture compression and SGI is NOT a graphics CARD COMPANY.

Neither is nVidia, they're a chipset company. :)
 
Doomtrooper said:
Proprietary extensions for a common function (GL_NV_vertex_program) is not texture compression
Trademarked instruction sets which require licenses to use are terribly common. By doing this, you prevent people from riding on your coattails and R&D. Are you going to lambast MIPS, Atmel, ARM, Intel, etc for protecting their instruction sets?

Doomtrooper said:
SGI is NOT a graphics CARD COMPANY.
SGI certainly is a graphics hardware company. Look at any of their workstations--they're tailored to graphical visualization.
 
First, I'm saying they had to make their own because NVIDIA didn't want them using theirs, which is the point of licensable technology--preventing your competitors to use your R&D money against you.

Next, did you even read the article? The extentions are different (the NV one and the EXT one) in their implementation, and they're discussion merging the two for OpenGL2.0's shader language. In the end, they both essentially expose the shading mechanisms of DX8, but the path to get there is differing.

Next, SGI's InfiniteReality graphics don't sound like the quadro to me--though its completely possible.
http://www.sgi.com/workstations/comparison.html (OK, the first two do)
http://www.sgi.com/Products/PDF/3226.pdf

HOWEVER, its completely irrelevant. When SGI started, they made their own graphics subsystems, they are essentially the fount of most knowledge with respect to 3d visualization. It originally was IrisGL, which ran solely on their Irix workstations. Then they made it OpenGL, and charged a pretty penny for licenses. They didn't do it for altruism.

Finally, look at http://www.oreillynet.com/pub/a/network/2000/06/23/magazine/opengl_intro.html.

Look on the web for a little bit of history and evidence for your positions. You might find yourself not being corrected so much if you'd do this first.
 
HOWEVER, its completely irrelevant. When SGI started, they made their own graphics subsystems, they are essentially the fount of most knowledge with respect to 3d visualization. It originally was IrisGL, which ran solely on their Irix workstations.

Which, co-incidentally, was licensed from 3Dlabs.
 
RussSchultz said:
First, I'm saying they had to make their own because NVIDIA didn't want them using theirs, which is the point of licensable technology--preventing your competitors to use your R&D money against you.

Yes, there are many things that are in the best self interest of a company that aren't in the best interest of the consumer or the industry. I believe the entire gist of the relevant portions of Doom's comments are that nVidia is establishing a track record of only self interest, which is not helping the graphics industry, nor consumers, as a whole unless you want to allow nVidia to provide all graphics cards and completely determine pricing (well, that would benefit consumers if they had infinite wallets). The parallels of Microsoft and RAMBUS come to mind.

The fact is, nVidia didn't invent the concept of shaders and shader programs, though they did invent their proprietary shader rendering assembly language, their implementation of this concept. Of course it made sense from the standpoint of nVidia's self interest to try and charge money for this implementation, but so does bank robbery if you manage not to get away with it. The problem is that OpenGL as it is NOW is an open standard for providing a common pool of ideas and implementation for all vendors to draw upon, and providing a common face for this for all developrs. nVidia seems to have established the distinction of having the most consistent track record of benefitting from others' contributions to this standard while seeking to charge for its own contributions...Cg seems another step in this path of one-way benefits.

Next, did you even read the article? The extentions are different (the NV one and the EXT one) in their implementation, and they're discussion merging the two for OpenGL2.0's shader language. In the end, they both essentially expose the shading mechanisms of DX8, but the path to get there is differing.

I believe the point is that one cannot be used without paying nVidia money, and the other CAN be used WITHOUT paying licensing to those who contributed it. The former is nVidia's standard method of operation now, and the latter is how the rest of the developers participate. I think the latter helps the consumer, vendors, and developers...the former, in contrast, only helps ONE vendor. Where would OpenGL be if all vendors practiced what nVidia does? nVidia is trying to leverage a direct stake of control and profit from the future of OpenGL...in effect hijack the future development of OpenGL and direct it towards its own destination (=$$$ for nVidia alone)...the problem other vendors may have with this is the foundation of contributions of other vendors on which these proprietary shader languages will rest.
 
Did YOU even read the article?

They got off on a bad foot with some of the legal positioning on NV_vertex_program, and would now like to offer the extension to the ARB free of any encumbrances or conditions, in contrast to the previous position.

Bimal asked if the previous constraint on adopting the extension only without changes still applied. David says that as we move away from features to programmability, NVIDIA is focusing on the processor instruction set. They'd resist changes which made the extension incapable of running on their platform, but would be more receptive to other changes. Going forward, use will be made less restrictive; IP does underlie the extension, but NVIDIA would like to license it freely and openly, without restrictions.

And this is from nearly a year ago.
 
I'm puzzled as to how you highlight text that points out how nVidia tried to act before, and makes clear how not trying to charge other vendors for licensing is in contratst to past behavior, and ask me if I've read the article. I make reference to track record, and say that they are trying to do the same thing they have done in the past, but now with Cg...it makes me question whether you really gave my post a good read. Perhaps I misused my tenses in discussing what nVidia has done and is now trying to do again with Cg, but as I reread I fail to notice where, and can't credit that your highlights contradict anything I stated. All it seems to do is point out that nVidia realizes they misstepped, yet is trying to do exactly the same thing again with their positioning of Cg and their requirement of licensing fees to offer a Cg implementation.
 
demalion said:
I'm puzzled as to how you highlight text that points out how nVidia tried to act before, and makes clear how not trying to charge other vendors for licensing is in contratst to past behavior, and ask me if I've read the article. I make reference to track record, and say that they are trying to do the same thing they have done in the past, but now with Cg...it makes me question whether you really gave my post a good read. Perhaps I misused my tenses in discussing what nVidia has done and is now trying to do again with Cg, but as I reread I fail to notice where, and can't credit that your highlights contradict anything I stated. All it seems to do is point out that nVidia realizes they misstepped, yet is trying to do exactly the same thing again with their positioning of Cg and their requirement of licensing fees to offer a Cg implementation.

How many companies in the past tried to sell something and then when they found out the market reaction were forced to give it away for free?

How many of your corporate heros that you cheerlead here were trying to license and maximize their IP a few years ago? Should we look at their past behavior as well before they were, "reformed" by the goodness of their hearts ( failure in the market place )?


Boy oh boy, Netscape was such a magnaminous company for giving away the source to their browser and going totally open. They are the heros of the open source world now. Microsoft is EVIL, Netscape is golden. But they only reformed after they tried to sell their browser for $30+ a pop and their application server for thousands, and MS completely crushed them by giving both away for free.

And of course, 3DLabs does not exist to make a profit, but to promote the general welfare of 3D apis. Oh no, they never tried to force IP licenses down anyones throats for an API.
 
demalion said:
All it seems to do is point out that nVidia realizes they misstepped, yet is trying to do exactly the same thing again with their positioning of Cg and their requirement of licensing fees to offer a Cg implementation.

Huh? Other vendors can still create their own compilers because of the open-source parser code and the public Cg language specification. It's true that the code generation isn't open-source but I cant really see why that should pose a huge problem for developers unless ATI, Matrox, 3Dlabs and others decides not to make that darn profile.

Remember that application developers can choose to provide assembly code for certain platforms (rather than work at the Cg level) according to Kurt Akeley. He goes on to state:

It's up to the application vendors to choose whether to pre-compile or trust the run-time compiler. In the current run-time the application specifies what compiler to use, so the application might query the system to see what compilers are present and use a particular compiler.

I tend to look at Cg this way: I love using Dreamweaver to make a website because I didn't bother learning everything about HTML, javascript and what not. In Cg the link between the Cg language (public) and the assembly code (public) is the code generation which isn't public (just as Dreamweaver works just great without me knowing its code generator).

It's a bloody tool for shaders - not a language that is impossed upon everybody in the World to speak solely from now on.
 
Back
Top