🔒 Locked

Hardware Acceleration

Started by typewriter Jan 19, 2005 at 3:11 PM 2 replies 450+ views

Original Post

Jan 19, 2005 at 3:11 PM

Greetings, Does the programmer have to make special provisions for cards which support different areas of hardware acceleration? Or does the video card simply 'automatically' take over in areas that it can? For example if I created a game on a machine with no opengl hardware accel, then moved it to a top of the line card, would the card just 'know' what to do, and automatically enable hardware acceleration for things like geometry, lighting or textures? How does this work? thanks a bunch for any links or insight into this mechanism

hplus0603

Jan 19, 2005 at 3:19 PM

In the case of OpenGL, yes, the card will accelerate the parts it can, out of the API you're using. Unfortunately, the linkable API for OpenGL only does version 1.1, so you'll want to use OpenGL extensions for the more exciting things, and those extensions wouldn't be available on cards that don't support the hardware for them.

On DirectX, you have to check a bunch of capability bits for what's actually supported or not; if you call things that aren't supported, it won't work right (and may even crash).

typewriter

Jan 19, 2005 at 4:53 PM

I realized that I never mentioned that my question was about openGL specifically.

These two replys seem to conflict with each other. From further reading it seems that SOME things will be automatically handled by capable hardware..

From reading about reflections in Mark Kilgard's article (http://www.opengl.org/resources/code/rendering/mjktips/Reflect.html), he alludes to the notion that his reflection technique will be automatically accelerated:

"And we've done it in a way that is completely hardware acceleratable."

So it seems that, at least for drawing geometry, the hardware will automatically accelerate it.

Topic Locked

This topic has been locked by a moderator. New replies are not allowed.