What is the deal. I was hoping that the new Detonators would fix the speed issue with render_to_texture/Pbuffers. It doesn’t it. It is faster than before. WIth the 80’s it was faster to do the copy your self. Binding the pbuffer was 3 to 4 times slower. NOw with the new 90 Dets. its faster but its still about 1 to 2 times slower. ITs better but i swear the gl driver is still doing a copy and The way its doing the copy from the p-buffer is still slower than the application doing the copy.
I was under the impression that render_to_texture purpose was to save the unnecessary copy and truely and honestly rendering right to the bloody texture. Even ATI’s implementation does actually render to the texture, they have the extra copy extension to do that.
So what is the point of render_to_texture if it cleary doesn’t actually render to texture?
Devulon