Archived

This topic is now archived and is closed to further replies.

wolfman8k

How does hardware rendering work?

Recommended Posts

wolfman8k    140
I started learning OpenGL a while ago. It''s pretty cool having two pieces of hardware working in paralel to boost speed: the cpu and graphics card. I''m just having trouble understanding exactly how it works. How do they split the work up? I know that the graphics card does the rendering, but what happens when it''s rendering? Does the cpu just continue through the code? But what if there is math involved in the rendering code(glVertex3f(5, 2, (x+2)/5) Does anyone know of an an article that discusses this? Thanks

Share this post


Link to post
Share on other sites
tcs    122
The CPU doesn''t render anything. If you want realtime FPS, you have to do everything in hardware. You hand your data to OpenGL and OpenGL is a layer above the gfx card driver. You shouldn''t care about how the driver (synchronous/asynchronous) passes the data to the card. On non T&L cards only the last steps of the pipeline are perfomed on the gfx card, because the CPU has to handle all tranformation and interpolation math for T&L.

"glVertex3f(5, 2, (x+2)/5" - Is this really a question ??? Before you can make a function call you have to have all parameter on the stack... And what does mean "math involved". Do you think rendering happens in some magical way without math ? ;-
)


Tim

--------------------------
glvelocity.gamedev.net
www.gamedev.net/hosted/glvelocity

Share this post


Link to post
Share on other sites
Mezz    571
As I understand it, hardware rendering is the process of using dedicated chips on the graphics card to perform operations that, under other circumstances, the CPU would have performed.
The CPU still needs to issue instructions to the graphics card, and get the data (geometry information) to the card, but once it is there the graphics card hardware takes over and leaves the CPU to do other things, like AI or physics.

-Mezz

Share this post


Link to post
Share on other sites
tcs    122
It depends... On T&L cards even the geometrie processing can be done on the card. How much depends on the T&L implementation. A GeForce can store all needed informations for rendering on the board and can render a full scene out of the local memory. I don''t know how much CPU work is needed for the card to render the scene, but it should be very low.

Tim

--------------------------
glvelocity.gamedev.net
www.gamedev.net/hosted/glvelocity

Share this post


Link to post
Share on other sites