nvidia, ati and opengl

[kryton9]

I was wondering about nvidia and ati. In studying opengl, you see functions that are ports from nvidia code or from ati code. One nvparse routine says it works great for NV20 based video cards. Does this mean you have to pick or write code to support both vendors? Or are these just special features you can tap into for custom apps for that card only?

How do we know what will work with what?

[Michael Hartlef]

As Petr mentioned, TBGL uses OpenGL 1.1 . Even if he uses 1.2 there should be agian no problems. When he starts to use Vendor specific OpenGL extentions, then he has to write the code for a specific vendor/card.

Generally with OpenGL it is wise to check which extensions are supported by a card and disable your feature then if not available or work around it.

[Petr Schreiber]

Hi kryton,

Mike is right. All what you can find in TBGL and thinBASIC openGL headers will work on any card.
But both ATi and nVidia offer their "personalised" version of headers ( even headers for GeForce 8800 for example ).
With such a modified versions you can get maxximum from determined card, but for example running nVidia advanced samples on my Radeon often results in bad output or I simply can't run the program.

According to some articles I read some years ago ( so may not be valid more ) you had to think deepely even design of shader programs, which were different/limited on determined architecture.

nVidia has nice program called nVemulate or similar, which allows you to get features of modern hardware on older GeForces. Of course those functions run on software, but at least you can play with them.

Do you have some special feature in mind you want to use ?


Thanks,
Petr

[kryton9]

Thanks Mike and Petr for the replies and explanations. I am enjoying reading and seeing what is out there for opengl. Still lots to learn, but it is fascinating and cool!!