Display of huge data

I have an extreme situation where i have to display 3D points (x,y,z) of the order of 90 millions in my viewport in Windowfit commands etc.

Does openGL as such has any limitations in handling data vis-a-vis the number?

What are Windowfit commands?

if i am not mistaken 90 millions points are something like 2GB only for double precision position data. that could be a problem!

by WindowFit, i mean fitting the entire data (90 million poits) into view port.

Nope. OpenGL is not a piece of hardware. It is an API to sending commands to the GPU.

Store your points into a few VBOs and render away. Make each VBO 4 MB.

by WindowFit, i mean fitting the entire data (90 million poits) into view port.

Does your viewport actually have 90 million pixels of resolution? If not, then you may want to consider some kind of culling to not render things that aren’t visible.

Also, there’s no way you’re going to fit 90 million positions onto the GPU. So you’re going to have to either use immediate mode or do some work dynamically transferring groups of pixels that you determine are visible.

How could one determine this?

Some explanation will be helpful for me to understand further.

Some explanation will be helpful for me to understand further.

1. putting it in a simple way if your resolution is 1024768 you have 1024768 pixels
2. if some of your pixels are behind other pixels you are not forced to draw them (or even store them)
3)as i told you, maybe your graphics card can’t handle 2GB of points data stored in one or multiple VBOs

and using immediate mode will take eons

How could one determine this?

I was being facetious. Unless you’re running at a resolution of 11250x8000, you don’t have 90 million pixels of display resolution. So, unless some of the pixels are translucent, you don’t need to draw that many pixels.

Ah but he did originally ask:

3D points (x,y,z) of the order of 90 millions

So even on a 1000 * 1000 (1 million pixel) display he could easily fit 90 million points given the fact that there is depth to consider.
So assuming he’s drawing these points with perspective projection he will definately need to attempt some type of optimisation where possible.
If these points need alpha/blending then depth sorting will be an expensive task.

This topic was automatically closed 183 days after the last reply. New replies are no longer allowed.