Difference between vertex_array and compiled_vertex_array

CVA''s prevent multiple transforms of the same vertices. Thats it.

Take this example, you need to render the same object several times, in a multipass fashion... Using standard vertex arrays, every vertex in the model is transformed at each pass ( and even more times if vertices are shared between glDrawElements calls ), which slows it all down. By Locking the array before rendering, the vertices are only transformed ONCE, and not in every glDrawElements call.

Death of one is a tragedy, death of a million is just a statistic.

thanx,

There are the locking functions in the compiled_vertex_array extension, which allow the array to get compiled...

when exactly do you use them for optimal results?

Depends what you''re doing...

For instance, say you are rendering a model several times, in a multipass rendering loop ( for whatever reason )... you could do something like:

LockArray( model_data )
RenderModel()
Change properties
RenderModel()
UnlockArray()
 

Also, say you have a terrain, and your rendering it using several triange strips, you''ll have duplicate vertices between the triangle strips, and so they''ll get transformed twice.. To prevent this, lock the entire array before hand, then call the individual glDrawElements calls for each of the strips.


Death of one is a tragedy, death of a million is just a statistic.

thanx for your help